In a world of rapid advancement and ever-shifting industry trends, continuous education and training are crucial for sustained employability. Accredify is developing a solution that can utilise technology to help corporates track the learning performance of their employees and devise necessary training interventions.

This project aims to provide a platform for our education customers to deliver a better (Learning & Development) L&D-as-a-Service experience to corporates. The platform will map courses to competency frameworks, assist corporates in identifying relevant courses for their learning & development needs and provide actionable insights with post-course evaluation. The project aims to tackle critical learners' pain points by identifying relevant courses and visualising how the additional knowledge or skills translate to improved performance.

Talent will work on post-commercialisation enhancements to the product, using implicit/explicit feedback and the user's profile to bring about more relevant recommendations.

Structural Health Monitoring (SHM) refers to the process of systematically analysing the data from existing infrastructures like skyscrapers and bridges, and extracting useful information on how they age. Such information enables predictive maintenance of critical infrastructures. We are seeking an enthusiastic individual to contribute to our SHM development efforts.

The apprentice will provide Research & Development assistance for the 4th Generation series of  Ackcio products.

Structural Health Monitoring (SHM) refers to the process of systematically analysing the data from existing infrastructures like skyscrapers and bridges, and extracting useful information on how they age. Such information enables predictive maintenance of critical infrastructures.

The apprentice will work closely with the Embedded and Hardware team, to develop and improve the current radio-frequency (RF) communication protocols and applications. They should take the initiative to see the product through the entire lifecycle from proof-of-concept to design and production.

Social media is ubiquitous and ever-evolving. Billions of people consume content on platforms like Instagram, TikTok, Twitter every day. Behind this enormous amount of data is a wealth of information about the world and its people, waiting to be unlocked.

Affable uses Deep Learning to analyse social media content at scale. They have analysed over 2 billion+ posts and 500 million+ public accounts via cutting edge NLP and Computer Vision models. To unlock the world's social media data is the mission of Affable, and the startup is looking for an apprentice to join them in that mission.

Affable uses Deep Learning to analyse social media content at scale. They have analysed over 2 billion+ posts and 500 million+ public accounts via cutting edge NLP and Computer Vision models. To unlock the world's social media data is the mission of Affable, and the startup is looking for an apprentice to join them in that mission.

This project is about researching, training and deploying new machine learning models to glean more insights from our dataset. This includes:

1. Location prediction of social media users via deep learning
2. Gender prediction of social media users via deep learning
3. Identification of same user across different social media platforms

The objective of this project is to understand the consumer sentiment for a given topic in the FMCG sector and generate engagement materials (ads, product concepts, etc.) to increase consumer engagement for a given topic.

In order to identify a consumer conversation precisely, they need to understand the semantics in the conversation, detect entities and topics, and understand the sentiment of those topics. Once these are being identified, they need to create conversation materials such as ads, tag lines, product concepts that are closer to the consumer conversations. They need to do this at scale to keep up with the fast-changing consumer world today. This will be challenging because of the multiple languages spoken in different regions.

The project will develop productivity-driven insights for the construction site team through deep-learning computer vision techniques.

CCTVs are present at almost all construction sites and are mainly used for manual progress monitoring or reviewing past footage if something goes wrong. The information will be tracked through the existing CCTVs at the sites to give the project team meaningful insights and knowledge to understand progress and identify potential risks early on. It is still a massive challenge for companies to understand their progress on a macro level as it is hard to make sense of all the information from so many different subcontractors. This results in a lack of visibility of issues on-site that can cause delays in concreting works due to the underutilisation of the assets.

The project attempts to build state-of-the-art AI solutions and automation systems for the built environment. The candidate will work alongside under guidance of a chief scientist to develop AI solutions in a few live sites for human activity recognition (HAR). We are looking for talent to speed up the development of appropriate AI models for more accurate defects and HAR with F1 recall greater than 90%. We are looking for ML engineers with a systematic mindset, willing to work at length to create solutions that never existed before.

Allay’s pipeline is built on its foundational platform technology that combines validated non-opioid analgesics and biopolymers for a new category of ultra-sustained therapeutics. As part of the pipeline program, the team is developing an injectables formulation that can be used for various clinical indications, again targeting for successful transition into post-surgical rehabilitation, reducing the need for constant re-dosing, clinical oversight and avoiding the drawbacks of systemic analgesics such as opioids.

As part of the pipeline development, the ATX30X project is currently working on local anaesthetic injectables. We aim to achieve long-lasting, tunable pain management while keeping our solutions minimally invasive. With a range of different form factors such as liquids, gels and pastes in our toolbox, we continue to improve and create new prototypes every day to bring patients relief. As an intern under the programme, the incumbent will be assisting the R&D engineer/chemist in analytical test method development for the pipeline project and any analytical-related testing using various equipment such as High-Performance Liquid Chromatography (HPLC), Gas Permeation Chromatography (GPC), UV-VIS Spectrometer, and Scanning Electron Microscopy (SEM).

The following are examples of the expected job scope:

•    Testing and characterisation of various prototypes, including in vitro elution testing and force measurements

•    Prototyping processes such as formulation and manufacturing

•    Participate in weekly team meetings.

•    Troubleshooting and calibration of laboratory equipment

•    Conducting validation tests on laboratory equipment

•    Other adhoc tasks as required by the team

Allay’s pipeline is built on its foundational platform technology that combines validated non-opioid analgesics and biopolymers for a new category of ultra-sustained therapeutics. As part of the pipeline program, the team is developing an injectables formulation that can be used for various clinical indications, again targeting for successful transition into post-surgical rehabilitation, reducing the need for constant re-dosing, clinical oversight and avoiding the drawbacks of systemic analgesics such as opioids.

As part of the pipeline development, the ATX30X project is currently working on local anaesthetic injectables. We aim to achieve long-lasting, tuneable pain management while keeping our solutions minimally invasive. With a range of different form factors such as liquids, gels and pastes in our toolbox, we continue to improve and create new prototypes every day to bring patients relief. As an intern under the programme, the incumbent will be assisting the R&D engineer/chemist in analytical test method development for the pipeline project and any analytical-related testing using various equipment such as High-Performance Liquid Chromatography (HPLC), Gas Permeation Chromatography (GPC), UV-VIS Spectrometer, and, Scanning Electron Microscopy (SEM).

The following are examples of the expected job scope:

•    Testing and characterisation of various prototypes, including in vitro elution testing and force measurements

•    Prototyping processes such as formulation and manufacturing

•    Participate in weekly team meetings.

•    Troubleshooting and calibration of laboratory equipment

•    Conducting validation tests on laboratory equipment.

•    Other ad hoc tasks as required by the team

Allay’s first product, ATX-101, is currently undergoing a global clinical trial for use in Total Knee Replacement surgery. It is designed to provide the weeks of pain management required for a successful transition into post-surgical rehabilitation, reducing the need for constant re-dosing, clinical oversight and avoiding the drawbacks of systemic analgesics such as opioids.

As part of the pipeline development, ATX10X project is focusing on providing pain relief to patients over a different duration for different clinical needs in comparison to our current pain relief investigational products (ATX101) which is currently undergoing clinical trials for Total Knee Replacement (TKR) surgery. As an intern under the programme, the incumbent will be assisting the R&D engineer/chemist in analytical test method development for the pipeline project and any analytical-related testing using various equipment such as High-Performance Liquid Chromatography (HPLC), Gas Permeation Chromatography (GPC), UV-VIS Spectrometer, and Scanning Electron Microscopy (SEM).

The following are examples of the expected job scope:

•    Work on the tunability of the drug release rate

•    To improve on the new manufacturing method used to manufacture the product while still providing the controlled release of the drug over a certain duration.

•    To alter the structural integrity of the product using different excipient

•    To find an alternate biodegradable polymer that has a faster resorption rate.

•    To explore the use of different API

Allozymes focus is on the rapid evolution of optimised enzymes with improved performance, revolutionising the way industries manufacture complex natural products. The company’s proprietary platform enables us to rapidly screen through thousands of enzyme variants for specific substrate activities, generating both genomic and analytical data. This data is large and requires us to establish efficient workflows for its management and usage. As a first step, Allozymes built a cloud-based data system that integrates into our current experimental data management system and allows streamlining data storage and analysis.

For the next step in this process, this data system will be used to analyse and build workflows for data processing, data analysis and to perform predictions using artificial intelligence approaches, such as Machine Learning and Deep Learning.

In this project, the apprentice will work on the development, implementation and deployment of artificial intelligence solutions aiming at improving Allozymes' current enzyme evolution platform. The apprentice will be trained to process and analyse data, implement machine learning algorithms, and deploy machine learning pipelines using real genomic and experimental data.

This project will require collaboration with the scientists and engineers in the company to achieve the proposed goals. The outcome of the project will enhance the performance of the Allozymes proprietary platform for engineering enzymes and microorganisms.

At Allozymes, we require the ability to read DNA molecules that are much larger than 150-250 bp (the range of short-read sequencing method), to allow us to analyse the distribution of mutations during the enzyme or strain engineering processes. This helps us to detect the mutants that provide better performance of an enzyme than the original and enables us to reproduce it.

Although we have established some in-house sequencing capabilities for the sequencing of bulk selected droplets to accelerate our developments for enzyme and strain engineering, more work is required for the development of selected droplets sequencing in a plate-based format. The current project aims to utilize Nanopore technologies and a liquid handler to sequence the selected droplets that are separated individually into each well. By the use of a liquid handler, we can automate the tedious process of extracting and amplification the DNA, before combing the samples for Nanopore sequencing preparation, to allow us to fine-tune the enzyme and its engineering process. In addition, Allozymes is using a proprietary droplet microfluidic system for enzyme engineering and subsequently, the DNA molecule is present in tiny amounts which further optimization is required to employ these sequencing technologies.

In this project, the apprentice will be extracting and preparing the samples for sequencing. The apprentice would be tasked to extract and handle multiple DNA samples, maintain the quality of the samples, and prepare the samples for sequencing. The apprentice will also be working with a liquid handler to streamline the sequencing process of processing hundreds of samples at a given time. This project will require collaboration with other scientists and engineers in the company. The outcome of the project will enhance the performance of the Allozymes proprietary enzyme engineering platform.

Carotenes are long-chain hydrocarbon molecules commonly found in the pigments of plants. These molecules play an important role in the life chain, such as photosynthesis. Phytoene performs the role of UV protection, antioxidants, and anti-inflammatory, and the cosmetic industry has started benefiting from it in its skin care products.

However, producing it from natural sources is resource intensive and costly and it takes 10,000kg of tomato skin to extract about 3kg of phytoene. Additionally, customer behaviour has significantly changed, and they demand ingredients from natural sources, which makes chemical synthesis approach less attractive for the cosmetic companies. Therefore, there is a need for biosynthesis of phytoene and similar compounds.

In this project, the apprentice will employ enzyme and strain characterisation and optimisation techniques to identify suitable enzymes and improve the bioprocess yield. The apprentice will need to conduct screening and characterisation of enzymes, improve the yield of production and optimise the bioprocess conditions.

Atomionics works on quantum sensors based on cold atom interferometry, where atoms are cooled down to almost absolute zero using lasers. At such cold temperatures, these atoms start behaving like waves, which allows us to measure the forces acting on them precisely.

This opens up a wide range of applications including, but not limited to, paving a path in the energy transition to zero-carbon, exponentially improving the performance of inertial navigation systems, and transforming the global positioning system. At Atomionics, you will get the chance to prototype, test, refine, finalise, and produce your designs toward making our devices compact and robust for field deployment. The only constraints are power and your imagination.

Our gravimeter has an electronics system comprising dozens of custom design PCBs used for everything, from power distribution, power sequencing, data signal processing, waveform generators, and filtering to radio frequency devices to control quantum components. This project is to develop and refine these PCBs to improve performance and reduce complexity, size, and cost.

Atomionics is working on quantum sensors, based on cold atom interferometry, where atoms are cooled down to almost absolute zero using lasers. At such cold temperatures, these atoms start behaving like waves, which allows us to measure the forces acting on them precisely.

This opens up a wide range of applications including, but not limited to, paving a path in the energy transition to zero-carbon, exponentially improving the performance of inertial navigation systems, and transforming the global positioning system. At Atomionics, you will get the chance to prototype, test, refine, finalise, and produce your designs toward making our devices compact and robust for field deployment with the only constraints being power and your imagination.

Our gravimeter comprises hundreds of components, including a titanium vacuum chamber, optical laser systems, and actuating mechanisms. This is combined with exotic materials, micron-scale tolerances, and complex assemblies. The project involves the continued mechanical engineering development of our machine, and its subsystems, to improve performance, reliability, quality, and reduce assembly costs through creative design and application of engineering principles.

Atomionics is working on quantum sensors, based on cold atom interferometry, where atoms are cooled down to almost absolute zero using lasers. At such cold temperatures, these atoms start behaving like waves, which allows us to measure the forces acting on them precisely.

This opens up a wide range of applications including, but not limited to, paving a path in the energy transition to zero-carbon, exponentially improving the performance of inertial navigation systems, and transforming the global positioning system. At Atomionics, you will get the chance to prototype, test, refine, finalise, and produce your designs toward making our devices compact and robust for field deployment with the only constraints being power and your imagination.

The project objective is to optimise SWaP (size, weight, and power) for a cold atom trap for atom interferometry application. The successful candidates will work with other senior physicists and mechanical engineers at Atomionics to design and build a miniaturised version of a magneto-optical trap (MOT), which can cool down atoms to micro-Kelvin temperature, taking into account required optical beam access for measurement. Design optimisation efforts will focus on the process of laser cooling of atoms, magnetic field calculations for an optimal trap configuration with anti-Helmholtz coils, and optical beam configuration for the measurement of gravity using atom interferometry in an ultra-high vacuum chamber.

Atomionics is working on quantum sensors, based on cold atom interferometry, where atoms are cooled down to almost absolute zero using lasers.

At Atomionics, you will get the chance to prototype, test, refine, finalise, and produce your designs toward making our devices compact and robust for field deployment with the only constraints being power and your imagination.

As part of the laser optical system of the quantum sensor, a DDS (Direct Digital Synthesis) is used as a waveform generator to control phase and frequency of the signal waves. This is controlled by an FPGA and requires firmware to be written in VHDL.

The apprentice will be a part of Augmentus's technical team responsible for developing a fully integrated and offline robot simulation environment that allows users to program industrial robotic systems in minutes instead of months. The apprentice will also develop robot programming features that are important for robotic development across various industrial applications.

Robot development is currently a fragmented and difficult process. This is particularly the case for dynamic automation development where AI and computer vision systems have to be integrated.

The apprentice will be a part of Augmentus's technical team responsible for developing a fully integrated and offline robot simulation environment that allows users to program industrial robotic systems in minutes instead of months. The apprentice will also develop robot programming features that are important for robotic development across various industrial applications.

BeeX is building the world's most advanced Hovering Autonomous Underwater Vehicles (HAUVs). The technology takes men and women out of dangerous environments, slashes carbon footprint, and ultimately creates a sustainable way of work.

BeeX was started as there was frustration surrounding the manual, dangerous and expensive methods of underwater work. Leveraging on 1 decade of R&D in underwater self-driving, the team focuses on helping partners achieve a transformative leap in conducting underwater operations. The company’s vehicles have worked in challenging conditions globally; providing insight on nearshore infrastructure like ports and vessels to further out on offshore wind parks.

As the Mechanical Intern, you will work on the design, production and testing of the mechanical systems that make up BeeX's flagship Hovering Autonomous Underwater Vehicle, and our upcoming vehicle design projects.

BeeX is building the world's most advanced Hovering Autonomous Underwater Vehicles (HAUVs). The technology takes men and women out of dangerous environments, slashes carbon footprint, and ultimately creates a sustainable way of work.

BeeX was started as there was frustration surrounding the manual, dangerous and expensive methods of underwater work. Leveraging on 1 decade of R&D in underwater self-driving, the team focuses on helping partners achieve a transformative leap in conducting underwater operations. The company’s vehicles have worked in challenging conditions globally; providing insight on nearshore infrastructure like ports and vessels to further out on offshore wind parks.

As the Electrical Intern, you will work on the design, production and testing of the electrical systems powering BeeX's flagship Hovering Autonomous Underwater Vehicle, and the team’s upcoming vehicle design projects.

Cellular agriculture has been one the important innovations in clean meat revolution and is one of the fastest growing sectors in biotechnology. Scaling up cells from tissue biopsies to 2D and 3D cultures for producing meat a.k.a. cultured meat is not so effective due to multiple factors. Culturing cells efficiently from tissue biopsies, lack of cost effective edible microcarriers for scaling up to produce cultured meat in the bioreactor has not been addressed to its entirety by cell based meat manufacturers.

Cellivate Technologies addresses these gaps through use of proprietary coating (ProCyte) on 2D surfaces as well as 3D edible microcarriers for producing cell-based meat products. Our ProCyte surfaces increases yield 2x to 4x and edible microcarrier prototypes have been well received by our customers. Pilot studies with our B2B customers have validated our technologies and we are currently under the process of scaling up to supply more customers.

Immune cells derived from the patient are isolated (Autologous Cells) have been genetically engineered and re-infused to patients with promising immunotherapies outcome. However, patient-derived cells are normally accompanied by a certain extent of functional impairment, senescence or exhaustion from prior treatment that led to a reduction in the potency of the products.  We are developing differentiated, off-the-shelf CAR-T cell therapies, derived from healthy donor T cells (Allogeneic Cells).  These genetically modified cells could serve patients immediately, and enable an overall efficient manufacturing process leading to reduction in cost of goods.

The key challenges faced in the development and generation of allogeneic CAR-products is the eliciting the graft-versus-host disease (GVHD) and allo-rejection in addition to durability of the products. The aim of this project is to address the above issues in CAR-T cell design and manufacturing. Specifically, we will explore various sources of immune cells including healthy donor derived Peripheral blood mononuclear cells, Cord blood mononuclear cells, Natural killer cells or Gamma-delta T cells, as well as different strategies that could potentially enhance the overall efficiencies of the allogeneic cellular platform.

Curium drives several technologies capable of performing rapid and automatic calibration of sensors in fully automated systems such as autonomous vehicles.

Currently, Curium is deploying increasingly advanced technologies into vehicles that are creeping up the level of autonomy from Level 2 to Level 4. This project aims to enhance these current systems and elevate these manufacturers to deploy Level 5 autonomous vehicles.

The current problem is that sensors can become miscalibrated for any number of environmental reasons. With these sensors providing erroneous reading, the decision-making algorithms will make incorrect decisions leading to injury or death to vehicle owners, occupants, or pedestrians.

By monitoring and automatically calibrating the sensors in real-time, we can improve uptime and reduce the likelihood of miscalibrations leading to a dangerous situation.

The apprentice will work with the Curium Development Team to develop novel techniques to measure, monitor, and adjust sensor data in near real-time for Automotive use-cases. The challenges are to problem-solve the issues related to modelling 3D environments in a highly efficient manner to leverage the limited computational resources that are typically embedded into onboard vehicle systems. 3D environments will be modelled primarily using Inertial Measurement Units (IMUs), LIDAR, RADAR, and Camera sensors and deployed on computational platforms such as (but not limited to) the NVIDIA platform.

DiMuto is a rising Agri Fintech company based in Singapore. We simplify every step of global AgriFood trade. From produce, to trade, to market, our AgriFood Trade Solutions help growers, exporters, and importers to trade efficiently with better visibility and finance.

One of the key parts of their solution is a Digitization Device (DACKY). The team is looking for someone to assist in helping to improve and refine the devices design, and manufacturing process, and also work together with the technical team to deploy AI to read real-time images on DACKY devices at our customer sites - basically how we can take better pictures, add just-in-time printing capabilities, deploy AI, deploy computer vision etc. The company is looking to make the DACKY more multi-purpose according to our various customer requirements so that it can better suit the operational demands and environment.

The apprentice will be developing numerical methods for Machine Learning and Optimisation Theory to enhance the current capabilities in Quantum Computing and contribute to Entropica Labs' core software technology and participate in customer projects.

The main goal of this project is to test advanced methods for Machine Learning to optimise Quantum computers' performances and benchmark Entropica's Optimisation Solvers by running benchmarking experiments using advanced annealing-based solvers.

The apprentice will learn and develop Quantum Computing approaches to real-world discrete optimisation problems. The main goal of the project is to bridge the gap between the most prevalent techniques in Optimisation, Machine Learning and near-term Quantum Computing.

The apprentice will be contributing to Entropica Labs' core software technology and participating in customer projects. You will analyse models and suggest implementations on Quantum computers while benchmarking the performances using standard and proprietary datasets. You will also get the opportunity to work directly with our customers, assisting them with testing and implementing optimisation workflows.

In this project, the apprentice will be part of the core development team and will focus on the algorithmic design, development and deployment of 2D/3D Computer Vision applications based on classical methods as well as Deep Learning. The apprentice will work on projects at various stages including development and implementation of future robotic perception capabilities, as well as troubleshooting and improvement of robotic perception pipelines in systems currently deployed at project site (80% onsite required).

Fabrica supports the automation of tile grouting where the robot can grout tile gaps and clean off excess grout including the edges autonomously in any given space with the goal of boosting productivity. Human intervention is only needed to reload the grout and water.

The project would revolve around real-world industry deployments taking our tile grouting robot from a TRL of 7 to 9. In addition to a tile grouting robot, we would also be working on other construction robots in alignment with our long-term vision. From August to December 2023, the team will be working on building reproducible and reliable PCBs for our robot. The company currently has its first version of fully manufactured PCBs for version 4 robots. This is a step up from the self-assembled PCBs in version 3. However, it should expect that there will be many bugs and if not, areas for improvements which would require redesigning. The redesign should not just improve the reliability but also the ease of manufacturing and assembly, and subsequent replacements of parts. As the grouting robot transitions into a more mature product, we need to think about user-friendliness.

Part of that for electronics design and assembly is to design for easy debugging and replacement of parts. This would include creating a simple user interface that could potentially generate error codes. The intern would also be involved in the initial electronics design and assembly for new robots, and potentially integrating tech stacks on existing robots that the team chooses to collaborate with.

Fabrica supports the automation of tile grouting where the robot can grout tile gaps and clean off excess grout including the edges autonomously in any given space with the goal of boosting productivity. Human intervention is only needed to reload the grout and water.

The project would revolve around real-world industry deployments taking our tile grouting robot from a TRL of 7 to 9. In addition to a tile grouting robot, the team would also be working on other construction robots in alignment with the company’s long-term vision. From August to December 2023, the team will be working on building a reproducible and reliable tile grouting robot for mass manufacturing. Fabrica’s robots are currently in the pilot production phase. The redesign should not just improve the reliability but also the ease of manufacturing and assembly, and subsequent replacements of parts. As the grouting robot transitions into a more mature product, there is a need to think about user-friendliness. Figuring out which part of the grouting robot requires the most human intervention and simplifying it would be key to boosting the productivity of the robots.

Talent would also be involved in the initial mechanical design and assembly for our new robots, and potentially integrating tech stacks on existing robots that the team chooses to collaborate with.

We are a fast-growing company, and we believe that the sustainable way to keep this growth is by educating online sellers via various digital content and free-to-use tools. At Konigle, we simplify the process of running an e-commerce store for the sellers and defining how to run and grow a store profitably. We are doing this by creating various productivity tools, automation and shortcuts for multiple e-commerce platforms, store builders, and many more.

Apollo is an eCommerce intelligence engine developed by the Konigle core team. It is the brain on which all of Konigle's intelligence decisions are made for our customers. Now, this is used by over 1,000+ customers in over 60+ countries. We're constantly adding capabilities to the brain. This involves coming up with insights such as forecasts, suggestions, reporting and many other functionalities. We want mentees to experience this firsthand, understand the system, and start adding functionality to it.

We are a fast-growing company, and we believe that the sustainable way to keep this growth is by educating online sellers via various digital content and free-to-use tools. At Konigle, we simplify the process of running an e-commerce store for the sellers and defining how to run and grow a store profitably. We are doing this by creating a suite of various productivity tools, automation and shortcuts for multiple e-commerce platforms, store builders, and many more.

Apollo is an eCommerce intelligence engine developed by the Konigle core team. It is the brain on which all of Konigle's intelligence decisions are made for our customers. Now, this is used by over 1,000+ customers in over 60+ countries. We're constantly adding capabilities to the brain. This involves coming up with insights such as forecasts, suggestions, reporting and many other functionalities. We want mentees to experience this firsthand, understand the system, and start adding functionality to it.

With the expansion of Nanolumi’s technology portfolio in lipid encapsulated fluorescence nano particles, the team aims to venture into offering novel imaging solutions to life sciences applications.

In this project, team members would be developing and evaluating the suitability of the new class of novel lipid encapsulated fluorescence nanoparticles in its key performance metrics against commercial products. As well as develop strong use cases for advance imaging techniques like Fluorescence lifetime imaging (FLIM) and two photon imaging.

Also, new surface functionality would be explored for the nanoparticles and evaluation of the materials with live cells required.

Everyday around the world people go blind from not having access to an eye specialist. Their technology brings eye screening to primary care, a paradigm shift needed to end preventable blindness. Using their AI technology, all doctors are empowered to grade fundus images and detect diabetic retinopathy, even without a specialist.

The apprentice will own a Deep Learning project from beginning to end. He/She will be responsible for defining an implementation strategy after the literature review, implementing the strategy in consultation with medical consultants and assisting in the deployment and validation of the solution.

MicroSec empowers organisations to take control over their environment through our unique approach of Security by Design at the Edge for IoT, IIoT and OT devices and networks, from proactively preventing attacks from the inside-out, stopping zero-day attacks on devices, preventing device and network intrusions, creating chains of trust between devices, and reducing your operational risks and costs starting from the device to the cloud.

MicroSec provides enterprise-grade PKI (Public Key Infrastructure) solutions to IoT devices. We also use our patented microcerts instead of traditional X509 certificates which are more amenable to being transported over low bandwidth protocols like BLE, NB-IoT etc.

However, some users may not be satisfied with this level of security with concerns about the power of quantum computing algorithms. In this case, one would need to rely on PQC (post-quantum cryptography) based techniques for key establishment and QKD (quantum key distribution) based techniques for key distribution.

The challenge in the application of PQC algorithms in the IoT space is not only the consideration of space and time requirements but also ensuring that the key and ciphertext sizes are kept within limits such that they can be effectively transported via low bandwidth communication protocols like LoRaWAN, BLE, and NB-IoT.

The objective then is to arrive at and implement a cryptographic scheme that has a reduction to a known NP-hard mathematical problem that is not solvable in polynomial time using any known quantum algorithms which are comparable to current ECC implementations on 32-bit ARM processors in terms of speed, memory requirements and code size with keys and ciphertext which can be effectively transported over low bandwidth protocols like LoRaWAN, Bluetooth, and NB-IoT.

MicroSec empowers organisations to take control over their environment through our unique approach of Security by Design at the Edge for IoT, IIoT and OT devices and networks. This is done by proactively preventing attacks from the inside-out, stopping zero-day attacks on devices, preventing device and network intrusions, creating chains of trust between devices, and reducing your operational risks and costs starting from the device to the cloud.

MicroSec provides enterprise-grade PKI (Public Key Infrastructure) solutions to IoT devices. We also use our patented microcerts instead of traditional X509 certificates which are more amenable to being transported over low bandwidth protocols like BLE, NB-IoT etc.

However, some users may not be satisfied with this level of security with concerns about the power of quantum computing algorithms. In this case, one would need to rely on PQC (post-quantum cryptography) based techniques for key establishment and QKD (quantum key distribution) based techniques for key distribution.

The challenge in applying PQC algorithms in the IoT space is not limited to considering space and time requirements. Key and ciphertext sizes must be kept within limits to effectively transport them via low bandwidth communication protocols like LoRaWAN, BLE, and NB-IoT.

The objective is to arrive at and implement a cryptographic scheme that has a reduction to a known NP-hard mathematical problem that is not solvable in polynomial time using any known quantum algorithms which are comparable to current ECC implementations on 32-bit ARM processors in terms of speed, memory requirements and code size with keys and ciphertext which can be effectively transported over low bandwidth protocols like LoRaWAN, Bluetooth, and NB-IoT.

In order to assist with the timeline, talent will need to have a good understanding of the specific experimental protocols and procedures (which will be trained by the assigned supervisor, such as the experimental protocols training, safety-related training sessions, SOP, etc). The most challenging part of our site is recoding the experimental data, and organising it in a neat manner as required by the ISO standard in terms of documentation after completing the experiments. This is because all the records will be required for the yearly ISO auditing purposes. So that the engineering team can focus on product development and design etc.

Talent can expect to learn hands-on lab skills, soft skills and industry standards and protocols to differentiate them from the rest of their peers. Take a look at what the real environment in industry labs are like to make informed decisions before joining industry. We warmly welcome the interns to join us after they graduate so that our talent pool will never dry out.

Reyal™ is a new, cloud-based, multi-layer solution with the technology and tools for companies to connect every physical product with a unique digital identity for secure product authentication & counterfeit protection. The Reyal™ solution combines the company's multi-disciplinary knowledge in advanced material science, photonics and electronics, as well as marketing.

With the solution expansion and product development, the authentication process and system need to be upgraded continuously, with the help of advanced signature materials development, deep learning, AI-based detection and analysis mechanism, data mining from the cloud database, pattern recognition and analysis through new methodology development (multispectral imaging, analysis software, and hardware integration, etc.).

A key area of the project would be to develop a new generation solution, such as but not limited to, unique anti counterfeit signature materials, AI-based detection and analysis mechanism, as well as a B2C user-friendly AI-based hardware/software (smartphone version reader/applications) which allows readout and analysis of unique and sophisticated spectral signature and/or images, or combination of any those technologies.

At the same time, within the Reyal broad project scope, a secondary task will be to improve and expand the Reyal cloud database in features, such as interface, dashboard, auto-monitoring, data auto-processing and reporting.

For the project, we aim to design and develop a 6 degrees of freedom motor controlled robotic system for X-ray and CT image-guided systems. We will first determine the parameters for the robotic application, then design the robotic mechanism, develop.

In the robotics design phase, it involves using 3D modelling software for conceptual and detailed design of robot components and motion control.For this project attachment, the selected candidate will get the opportunity to work with experienced mechanical engineers, software engineers and AI experts during the system integration.

In natural language processing, large language models, which use massive neural networks to model statistics of natural text, are redefining how companies think about search, chatbots, copywriting, text editing, code development, and more. Models like ChatGPT use massive transformer networks trained on large web text corpora to be able to respond to user queries via extending user-supplied text prompts with uncanny ability. These responses are based on learning natural statistical patterns in the training text and generating responses based on the probability assigned to each following word given some prefix text. The remarkable capabilities of these models have been unlocked by the ability to scale transformer language models to massive sizes using huge amounts of GPU computing and to train them on enormous text corpora scraped from the internet.

Much like natural language, proteins are sequences of amino acids that fold into three-dimensional structures to carry out most functions at the molecular level of life. Also, much like natural language, we now have enormous databases containing the amino acid sequences of natural, functional proteins. Although most of these proteins have not been characterised (we only know their sequences), it turns out that statistical analysis of just these sequences can reveal evolutionary pressures, and, therefore, structural, and functional characteristics.

Over the past few years, large-scale deep learning methods, like protein language models, have transformed our ability to understand and predict the structural and functional properties of proteins by learning from these evolutionary patterns. However, current protein language models and their extensions (e.g., AlphaFold2 or ESMfold) have only scratched the surface of what large protein language models can enable for protein design and optimisation.

We have developed large protein language models that enable functional, prompt-based protein generation.

The objective of this project is to work with our machine learning team to develop the next generation of protein language models and to integrate these into our OpenProtein.AI platform for solving function-driven protein design tasks.

We are a lithium-ion battery recycling start-up that uses our patented world’s first redox targeting electrochemical process to recycle batteries. We strive to push for a cleaner battery recycling environment and for a better place for future generations. We do this by transforming the recycling industry by reducing pollution and contributing to a cleaner battery recycling environment. With support Temasek Foundation, Momentum VC (SMRT) and local battery players, we are building our plant in Singapore.

We are a young and fast-growing start-up, looking for individuals who have the energy and willingness to push the boundaries of technologies. We are looking for electrochemical researchers who have experience in any of these areas; redox flow process, fuel cell, vanadium flow batteries, electrolysers.

This project comprises three main parts:

• Research and development of new materials for our proprietary electrochemical process for Lithium recycling. This involves research of new material compositions, the synthesis and preparation of these materials, and their characterisation.
• Testing the newly synthesized material compositions regarding their electrochemical performance with wet chemical and instrumental analysis methods (Setup of experiments, performance tests and post-analysis).
• Review and evaluation of the results, based on high quality experimental data obtained from wet chemical and instrumental analysis methods. The experimental results will be compiled in a report together with the experimental workflow to summarize and compare with existing data.shboard, auto-monitoring, data auto-processing and reporting.

Intensive farming and unregulated usage of antibiotics in agriculture has led to development of antibiotic resistance that spilled over to human population, creating healthcare burden worldwide.

Peptobiotics has identified this issue and aims to develop recombinant livestock feed additives as replacements for agricultural antibiotics and hormones.

Peptobiotics is a seed funded startup that is currently in the stage of scaling up our production based on promising antibiotic alternatives that we have discovered.

In this project, the apprentice will apply their synthetic biology and molecular biology knowledge to engineer our Anti-Microbial Peptide (AMP) producing strains to reach higher productivity.

As most of the microscopes are standalone machines and used by an operator in an offline condition, this project requires the microscopy to be connected online live and be operated remotely. The remote access and control can be through web pages or AR/VR systems. The talent will be involved in the IoT development of microscopy.

The objective is to build up the IoT ego system to prepare the microscopy for future Metaverse applications. It is not common in the industry to connect the microscopy or remotely access it through the internet. Synchronising the operation of the scope without crashing into the sample is one big challenge.

One key function of microscopy is acquiring images. This is very time-consuming, from image acquisition to the analysis of the images. We have already automated the process of image acquisition. But there are still many improvements in image processing, analysis, and decision-making. Therefore, deep learning, pattern matching, and AI will help reduce the processing time by cutting down the manual work. The challenge is to build a system that allows us to speed up the image processing and deep learning to speed up the process between different types of samples.

Leading the advanced optical industry, Phaos’ optical Microsphere Nanoscopy (ONM) technology has achieved a breakthrough in optical and electron microscopy. The resolution of an optical imaging system has a principal limitation to resolving two adjacent objects due to the physics of diffraction. With the control of light’s wavelength and assistance by the microsphere, OMN can resolve target features down to 137 nm, which is way below the physical diffraction limit of 200 nm under visible light. This is made possible using simulation software, DOE (diffractive optical element), and the optical system’s physical design.

This project is to conceptualise and design amicrosphere to assist microscopic products that will be able to resolve two adjacent objects up to the size of 50nm using simulation software, DOE and physical design of the optical system.

Polybee is building autonomous drone solutions for pollination in indoor greenhouse farming. We have developed a novel patent-pending method of pollination with drones called Aerodynamically Controlled Pollination. This is a bio-inspired approach; the turbulent wake of drones is optimized to vibrate the flowers at just the right frequencies, much like the bumblebees when they buzz-pollinate. With this method, we can pollinate tomatoes, pepper, eggplant, and strawberries: the biggest crops in the greenhouse sector. Our core pillars of technology are automation of micro-drones, computer vision for perception and analytics, and a software platform for deployment and visualization. In this project, you will be involved in development of computer vision pipelines to do crop measurements and liaise with our robotics and horticulture teams.

As a computer vision intern at Polybee, you will be working on developing and deploying deep learning and image processing pipelines for high throughput phenotyping. You will be using data from a multitude of sensors (stereo-camera/NIR/thermal/climate) to derive valuable insights on plant morphology, plant health and detect stress, diseases, etc. This involves translating cutting-edge research into cost-effective, scalable industry solutions. If working on interesting real-world problems in a fast-paced multi-disciplinary team excites you, Polybee wants to hear from you!

In this project, you will be involved in deploying autonomous micro-drones in greenhouses, and liaise with the computer vision and software teams.

As an aerial robotics apprentice at Polybee, you will find yourself in various dynamic and high-impact situations; one day you could be writing software to make a micro-drone fly itself to a flower and perform pollination, on another day you might be developing a mobile app to help their users interact with their robotic bees. You will have the opportunity to take ownership of your work and be involved in deploying your own contributions at their customers' sites.

With the current process of polymer development, development scientists draw on their material knowledge to formulate new compound recipes, mix small batches on trial and test and then repeat the loop until the desired material properties are achieved, resulting in countless development hours. Complex recipes can take more than a year to develop.

Polymerize enables R&D scientists to reduce the trial & error in formulation development with a complete product lifecycle management software that will allow companies to efficiently access/share data and maintain formulation workflow orders, applications recommendations, reporting and analytics.

The MLOps role is a challenging, steep, and hands-on role for individuals ready to take full responsibility for the AI/ML applications within the Polymerize platform. The goal is to ensure the platform runs seamlessly with maximum automation and minimum supervision. We plan to achieve this by building a highly scalable, consistent, and seamless workflow with strong technical expertise, the latest technologies and state-of-the-art algorithms for ML and AI. We work as a closely knit team with no hierarchy and a constant push for ownership and customer-centric delivery of responsibilities. 

This role is also about having a lot of fun brainstorming, sometimes profoundly enlightening discussions on common industry problems, technology limitations, domain-specific roadblocks, and applications of AI. We strongly vouch for flexibility, deep work, and a balance of learning something you don't know while teaching something you know for the team's collective growth.

This project will focus on biomechanical analysis for clinical applications, more precisely on the evaluation of the knee joint movement for a better understanding and diagnosis of knee joint injuries and pathologies.

GATOR by PreciX is used to collect 3D Knee kinematics data during various movement activities by patients. This time series data is multidimensional and complex which demands use of ML based techniques for analysis.

The main purpose of this study is to develop algorithms to analyse altered movement patterns in knee osteoarthritis (OA) patients. Understanding kinematic patterns in patients with functional impairments will help early diagnosis of OA and develop treatment options.

This will include identifying distinguishing features in the clinically referred movement exercises performed by the subjects and successful training machine learning models for diagnosis. Precix has a library of biomechanics datasets with different movement profiles to be used during the project.

PreciX is a MedTech startup focused on developing a novel wearable device for the knee joint. This device aims to provide a range of biomechanical measurements, by leveraging a modular and scalable multi-sensor platform. The clinician will use these measurements to accurately assess the functional performance of the knee joint during natural dynamic movement, thereby enabling a range of clinical applications such as screening, diagnosis support, and recovery tracking.

The project will focus on biomechanical analysis for clinical applications, more precisely on the evaluation of the knee joint movement for a better understanding and diagnosis of knee joint injuries and pathologies. The main purpose of this study is to develop algorithms to analyse altered movement patterns in ACL injured patients.

RoPlus aims to help local companies automate their production lines for more innovative and faster production processes by providing automated gripping solutions with computer visioning. Our research experience in robotics and computer vision allows us to provide feasible and effective solutions for automation challenges in different industries. Currently, we are expanding our team for our product launch.

The robotics engineer is responsible for the evaluation of our gripper products with industrial partners, conducting durability tests to come to our product specifications, and supporting product developments for customisation projects at RoPlus. The apprentice should have experience in using open-source object detection algorithms or ROS motion planning API, and in dealing with industrial robotics systems. Fresh graduates with robotics experience are welcome to apply for the role as well.

RoPlus aims to help local companies automate their production lines for more innovative and faster production processes by providing automated gripping solutions with computer visioning. For the past three years, our R&D team has been focusing on developing soft grippers, exploring various automated solutions for production lines or retail distribution centres, and integrating computer vision in pick and place tasks. Our research experience in robotics and computer vision allows us to provide feasible and effective solutions for automation challenges in different industries.

RoPlus provides intelligent reconfigurable gripping solutions to increase packaging efficiency. Our technology is expected to impact a range of industries, including but not limited to food assembly, vertical farming and fast-moving consumer goods packaging. Currently, we are expanding our team for our product launch. The apprentice will be a part of RoPlus's technical team responsible for developing standard protocols for validating product characters and providing proposals on higher repeatability results and quality management. In the end, we aim to have the apprentice monitoring our product performance prior to the official delivery to the client or to the next stage of integration.

To launch a constellation of satellites to enable worldwide entanglement distribution, serving as the communication backbone for the quantum internet.

The apprentice will be supporting the R&D of space-based Quantum Entanglement instrumentation at SpeQtral, including entanglement generation onboard a nanosatellite and optimising quantum instrumentation. This can cover both hardware and software development efforts.

SpeQtral is commercialising quantum communications to prepare the world for the quantum future. Think quantum space lasers, entanglement and military level secure communications. As part of our mission to transform the world's networks, we plan to launch a constellation of satellites to enable world-wide entanglement distribution. The earliest application of this was the highly secure delivery of encryption keys to facilitate secret communications.

The delivery of secret keys via quantum entanglement (or quantum key distribution, QKD) originates at a physical level, where light particles travel from one location to another and are detected by extremely sensitive detectors. However, once the light particles are exchanged and detected by the communicating parties, the extraction of the information that makes up the secret key is entirely reliant on software methods. In the QKD field, the necessary software to deliver the secret keys is known as the QKD software stack. From signal processing challenges (i.e. detection of correlation peaks in high-loss environments) to optimisation of security (i.e. implementation of error correction codes or privacy amplification) the QKD stack presents several challenges that need to be addressed in order to bring entanglement-based QKD from academic levels to commercial maturity.

To launch a constellation of satellites to enable worldwide entanglement distribution, serving as the communication backbone for the quantum internet.

The apprentice will be supporting the R&D of space-based Quantum Entanglement instrumentation at SpeQtral, including entanglement generation onboard a nanosatellite and optimising quantum instrumentation. This can cover both hardware and software development efforts.

SunGreenH2 manufactures core components for electrolyser cells, stacks and systems, enabling our customers to produce affordable green hydrogen. We use our platform technology to incorporate proprietary advanced nanostructured materials into electrolyser components. Electrolysers made with our components dramatically increase production and decrease energy consumption with 30x reduction in the use of expensive platinum group metals.‍

Our novel, breakthrough technology results from over 10 years of cutting-edge research and innovation in electrochemistry and nanotechnology for renewable energy generation. We have a vision for zero carbon, low-cost, green hydrogen available globally at scale.

Tigtag's mobile application supports drawing squares and polygons around objects in images to label them. There is a current challenge to do polygonal bounding boxes when points are added one at a time around the border of the object.

In this project, apprentices will explore developing a new feature where the user highlights a square area around an object, and a resulting polygon around the object is automatically generated with points on the image.

This will require an object detection algorithm to be built with points plotted with high level of accuracy around the object. This feature will be game changing for the company as we are a mobile first company with no other players in the market is currently doing this to achieve economy of scale for drawing polygonal bounding box to power computer vision projects.

High cost of production is one of the key bottlenecks in the cultivated meat industry. This project aims to address this problem through identifying suitable easily accessible and affordable plant-based scaffolds, recognizing scale-up parameters using an inexpensive mini bio-reactor and optimizing media feeding strategy.

Key areas within project include characterisation of the mini bioreactor process parameters, screening multiple plant scaffolds for use as inexpensive easily available scaffolds for long term culturing of fish cells; Culturing of fish cells in scaffolds for long term in custom made mini bioreactors and reducing cost of scaffolds for long term fish cell culturing.

Other areas of focus include protecting shear sensitive cells from hydrodynamic shear by employing 3D scaffolds for cell proliferation and shear protection; also, bio-process scale up through characterisation and operation of an inexpensive mini bio-reactor and achieving cost reduction through optimisation of media feed strategy.

Echocardiography (ultrasound of the heart) is the first test of choice for Cardiologists since, compared to other imaging methods, as it is less expensive, more mobile, non-invasive and radiation-free. However, the image quality is lower than other techniques like MRIs and CT scans.

US2.ai is trying to bridge the gap by developing software to bring AI-enhanced imaging resources to everyone. The project consists of developing AI tools to democratise echo, the most commonly used tool for detecting heart risk.

Vebits is a leading provider of intelligent infrastructure inspection solutions. Our cutting-edge product, RoadVisionPro, has gained significant recognition in recent news coverage by esteemed outlets such as CNA, Channel 8, and the Straits times. RoadVisionPro leverages artificial intelligence to automate the inspection of infrastructures and generate comprehensive reports embedded with professional metadata, thereby facilitating efficient maintenance practices.

The team is currently seeking a qualified candidate to join our team and contribute to the advancement of this project. In this role, talent will be responsible for enhancing our existing solution to ensure scalability for a broader range of customers and applications, all while upholding the utmost standards of security and privacy.

Requirements for this position include a strong grasp of AI technologies, proficiency in model analysis and improvement, expertise in deploying solutions for multiple customers, and a meticulous approach to maintaining service excellence. The ideal candidate would also possess a deep understanding of scalable infrastructure systems, privacy concerns, and security protocols. As part of the project, talent would be developing and deploying AI solutions, including data analytics, machine learning, OpenCV and deep learning. Talent would also be responsible for designing, developing, testing, and implementing AI systems that solve complex business problems.

Venti is developing self-driving container trucks, aka Autonomous Prime Movers (APMs) to move shipping containers between cranes within the PSA Singapore port. The project aims to alleviate the strains in the supply chain by addressing manpower shortages and further improving operational efficiencies. The port offers unique challenges for the deployment of self-driving vehicles such as high environment change (variable stock of stacked shipping containers, moving cranes), high occlusions from large actors and structures, all-weather operational demands, tight positional tolerances for interfacing with cranes, interfacing with port specific infrastructure, specialised port traffic patterns with corresponding traffic rules, and control of long articulated vehicles (truck with trailer) under variable loading up to a maximum of 65 metric tonnes. Other problems are common to urban traffic scenarios, such as mixed traffic (human and robot) driver navigation of intersections and lane changes with travelling speeds of up to 40 kph, which demands APM perception range of 150m.

The wharf is designated as the few lanes alongside the vessels (large container transport ships), serviced by quay cranes. In this area, there is an absence of static global features, as the large quay cranes move on rails. There are multiple bi-directional lanes (to service container-facing requirements on refrigerated containers), and driving routes between the wharf and the yard (the remainder of the port) are made across an unmarked area called the back reach. The lift height of the cranes is also higher than in the yard, and the APMs must obey port rules to yield in the presence of suspended loads. Additionally, there will be many pedestrians in the wharf area to attach or remove small mechanisms called twist locks from the corners of the containers (these mechanisms are used for alignment and lashing to the vessels).

The focus of the graduate student research will be on the localisation problem for entry, exit and traversals within the wharf area of the port. This is a unique challenge due to

a) lack of static features on the global map,

b) demand for precise alignment to quay cranes for mounting/offloading containers,

c) high occlusion due to large structures in the surroundings, and

d) potential localisation jump issues when transitioning back to the yard, ie switching from a relative position tracking or dead reckoning back to global position feature matching methods.

This may involve tracking of the vehicle’s position relative to specialised port features and/or experimentation with new sensor modalities. Minor infrastructure changes can also be explored if and as needed.

Venti is developing self-driving container trucks, aka Autonomous Prime Movers (APMs) to move shipping containers between cranes within the PSA Singapore port. The project aims to alleviate the strains in the supply chain by addressing manpower shortages and further improving operational efficiencies. The port offers unique challenges for the deployment of self-driving vehicles such as high environment change (variable stock of stacked shipping containers, moving cranes), high occlusions from large actors and structures, all-weather operational demands, tight positional tolerances for interfacing with cranes, interfacing with port specific infrastructure, specialised port traffic patterns with corresponding traffic rules, and control of long articulated vehicles (truck with trailer) under variable loading up to a maximum of 65 metric tonnes. Other problems are common to urban traffic scenarios, such as mixed traffic (human and robot) driver navigation of intersections and lane changes with travelling speeds of up to 40 kph, which demands APM perception range of 150m.

The wharf is designated as the few lanes alongside the vessels (large container transport ships), serviced by quay cranes. In this area, there is an absence of static global features, as the large quay cranes move on rails. There are multiple bi-directional lanes (to service container-facing requirements on refrigerated containers) and driving routes between the wharf and the yard (the remainder of the port) are made across an unmarked area called the back reach. The lift height of the cranes is also higher than in the yard, and the APMs must obey port rules to yield in the presence of suspended loads. Additionally, there will be many pedestrians in the wharf area to attach or remove small mechanisms called twistlocks from the corners of the containers (these mechanisms are used for alignment and lashing to the vessels).

The focus of the graduate student research will be on the perception problem for entry, exit and traversals within the wharf area of the port. More specifically, the APM will be required to:

a) accurately track the positions of quay cranes, prime movers, pedestrians, and other specialised port vehicles in the area, and

b) to detect the presence of any suspended load that the quay cranes may be handling, where typical loads are shipping containers or vessel hatch covers of various sizes.

This is a unique challenge due to:

a) lack of public sensor datasets for training over port equipment/vehicles,

b) demand for precise alignment to quay cranes for mounting/offloading containers,

c) high occlusion due to large structures in the surroundings, and

d) high lift heights of the quay cranes upwards of 50m.

This may involve experimentation with new sensor layouts and/or new sensor modalities.

Venti is developing self-driving container trucks, aka Autonomous Prime Movers (APMs) to move shipping containers between cranes within the PSA Singapore port. The project aims to alleviate the strains in the supply chain by addressing manpower shortages and further improving operational efficiencies. The port offers unique challenges for the deployment of self-driving vehicles such as high environment change (variable stock of stacked shipping containers, moving cranes), high occlusions from large actors and structures, all-weather operational demands, tight positional tolerances for interfacing with cranes, interfacing with port specific infrastructure, specialised port traffic patterns with corresponding traffic rules, and control of long articulated vehicles (truck with trailer) under variable loading up to a maximum of 65 metric tonnes. Other problems are common to urban traffic scenarios, such as mixed traffic (human and robot) driver navigation of intersections and lane changes with travelling speeds of up to 40 kph, which demands APM perception range of 150m.

The wharf is designated as the few lanes alongside the vessels (large container transport ships), serviced by quay cranes. In this area, there is an absence of static global features, as the large quay cranes move on rails. There are multiple bi-directional lanes (to service container-facing requirements on refrigerated containers), and driving routes between the wharf and the yard (the remainder of the port) are made across an unmarked area called the back reach. The lift height of the cranes is also higher than in the yard, and the APMs must obey port rules to yield in the presence of suspended loads. Additionally, there will be many pedestrians in the wharf area to attach or remove small mechanisms called twistlocks from the corners of the containers (these mechanisms are used for alignment and lashing to the vessels).

The focus of the project will be on the vehicle motion control problem for entry, exit and traversals within the wharf area of the port.

This is a unique challenge due to:

a) the broad range of loading conditions in the vehicle’s attached trailer, from unladen to laden with 65 metric tonnes,

b) tight positional tolerances needed for interaction with cranes offloading or mounting the containers, and

c) passing through tight spaces when navigating past traffic congestion.

Additionally, some port layout variants will require accurate reversing with the trailer attached.

Venti is developing self-driving container trucks, aka Autonomous Prime Movers (APMs) to move shipping containers between cranes within the PSA Singapore port. The project aims to alleviate the strains in the supply chain by addressing manpower shortages and further improving operational efficiencies. The port offers unique challenges for the deployment of self-driving vehicles such as high environment change (variable stock of stacked shipping containers, moving cranes), high occlusions from large actors and structures, all-weather operational demands, tight positional tolerances for interfacing with cranes, interfacing with port specific infrastructure, specialised port traffic patterns with corresponding traffic rules, and control of long articulated vehicles (truck with trailer) under variable loading up to a maximum of 65 metric tonnes. Other problems are common to urban traffic scenarios, such as mixed traffic (human and robot) driver navigation of intersections and lane changes with travelling speeds of up to 40 kph, which demands APM perception range of 150m.

In addition to considering the safety and performance of a single APM, we also must ensure a safe and performant operation of a fleet, supported by remote operations and reliant on external systems such as wireless network infrastructure. Complimenting the autonomous system with remote operations can ensure more uptime and safe handling of exceptional cases, but we must further ensure that the functional safety design brings the vehicle autonomy to a safe state for handover to the remote operator, so that the remote operator can resolve the situation safely.  Loss of network connectivity or high latencies could further compromise the remote operator's ability to control the APM, hence remote operations functions and fail-safes need to be designed with such limitations in mind. Designing failsafe manoeuvres will require a deep understanding of the multi-objective functions reflected in system requirements, where sometimes vague text definitions such as drive defensively, may need to be translated into code that the robots can interpret.

The focus of the project will be on systems design for the functional safety of a robot and human combined system of multiple APMs and remote operators connected over a wireless network. This may include following steps of ISO standards such as hazard analysis and risk assessment, failure modes and effect analysis, the definition of critical safety and performance metrics for triggering failsafe actions (safe stop, request intervention, etc), and quantification of residual risks. The scope may further include identifying patterns in network degradation, where if network degradations are predictable due to effects of weather, shipping container movements, vehicle speed, or other observable environmental factors, fleet management mitigations can also be designed (ie detour to avoid network dead zones).a

Venti is developing self-driving container trucks, aka Autonomous Prime Movers (APMs) to move shipping containers between cranes within the PSA Singapore port. The project aims to alleviate the strains in the supply chain by addressing manpower shortages and further improving operational efficiencies. The port offers unique challenges for the deployment of self-driving vehicles such as high environment change (variable stock of stacked shipping containers, moving cranes), high occlusions from large actors and structures, all-weather operational demands, tight positional tolerances for interfacing with cranes, interfacing with port specific infrastructure, specialised port traffic patterns with corresponding traffic rules, and control of long articulated vehicles (truck with trailer) under variable loading up to a maximum of 65 metric tonnes. Other problems are common to urban traffic scenarios, such as mixed traffic (human and robot) driver navigation of intersections and lane changes with travelling speeds of up to 40 kph, which demands APM perception range of 150m.

The wharf is designated as the few lanes alongside the vessels (large container transport ships), serviced by quay cranes. In this area, there is an absence of static global features, as the large quay cranes move on rails. There are multiple bi-directional lanes (to service container-facing requirements on refrigerated containers), and driving routes between the wharf and the yard (the remainder of the port) are made across an unmarked area called the back reach. The lift height of the cranes is also higher than in the yard, and the APMs must obey port rules to yield in the presence of suspended loads. Additionally, there will be many pedestrians in the wharf area to attach or remove small mechanisms called twist locks from the corners of the containers (these mechanisms are used for alignment and lashing to the vessels).

The focus of the project will be on the navigation problem for entry, exit and traversals within the wharf area of the port. This is a unique challenge due to the port's guiding rules for navigation to:

a) position the vehicle paths relative to semi-static quay cranes,

b) comply with sequences of queuing/holding instructions, and

c) coordinate with the other nearby vehicle bidirectional traffic patterns.

The APM will also need to adjust paths to circumvent traffic congestion or crane maintenance/servicing zones.

3D occupancy perception, also known as stereo depth perception, is a fundamental building block of spatial perception. The underlying principle is called triangulation, i.e., finding matched points between two adjacent overlapping images captured by a stereo camera, and calculating the disparities between the matched points to derive depth information.

This project is established for a key purpose of driving products that will democratise spatial perception of future autonomous and tele-operated robotic systems. Vilota specialises in vision perception solutions for mobile robots to achieve autonomous navigating capabilities in GPS-denied environments.

The target and emphasis is to grow a suite of vision sensors and further develop an advanced vision perception software stack.

By using state-of-the-art dense depth estimation methods, the primary aim of this project is to use vision-based perception to produce images of significantly higher resolution in comparison to competing technologies using multiple cameras.

Robotics visual perception can generate two important types of information - geometry and semantics. The former has been well developed by several major players in the market, while the latter poses a higher technical barrier and therefore has become the focus of this job position.

This project is established for a key purpose of driving products that will democratise spatial perception of future autonomous and tele-operated robotic systems. Vilota specialises in vision perception solutions for mobile robots to achieve autonomous navigating capabilities in GPS-denied environments.

The target and emphasis are to grow a suite of vision sensors and further develop an advanced vision perception software stack.

To achieve scalable autonomous applications, we currently develop a set of algorithms to recognize navigational-related features (like road signs and markings) common obstacles (like windows, doors, trees, and other mobile robots), and to further reason about the interrelationships of the objects in an image.

In this project, candidate will build upon the existing deep-learning frameworks currently being used by the industry.

Visual inertial odometry is an uprising approach for robotics localisation as an alternative to the more mature and conventional LiDAR-based localisation. Inevitably, a shortcoming of vision perception is that it has a shorter perception than LiDAR’s, making it often less reliable in vast areas. However, visual perception still appeals to a large audience of technical users because of image sequences contain huge amount of information which have many other potential uses.

This project is established for a specialised field that focuses on vision perception solutions for mobile robots to achieve autonomous navigation capabilities in GPS-denied environments.

The target and emphasis are to grow a suite of lightweight vision sensors and further develop an advanced vision perception software stack.

Hence, in this project, the team’s primary aim is to fuse object instance detection results with the vision-based odometry to achieve optimal localisation with little or no drift in commercial settings.

WaveScan constantly advances its technology with their expert team of scientists and engineers at WaveScan and at A*STAR.

As their technology advances, they deal with a wide array of composite and building materials that require robust signal post-processing and filtering algorithms to ensure the 3D holographic image produced has high resolution and is accurate and precise. As they aim to integrate the scanner onto platforms with large amounts of jitter such as drones, their signal processing algorithms are key to ensuring the image is clean and precise.

WaveScan specialises in the R&D, manufacturing and commercialisation of radar imaging technology for non-destructive testing (NDT) applications.

System integration is key to their solution, allowing customers to quickly deploy their technology through full-scale automation for their diverse inspection needs. Some automation platforms that their scanners work with include robotic arms, robotic crawlers, stationary mounting platforms, UGVs and drones.

They are test-bedding their technology with multiple infrastructure stakeholders such as HDB, JTC, and private residential developers in Singapore. They are also scaling up overseas markets in Japan, the U.K. and Europe.

This project touches on three key themes - 
The first is the design of high throughput synthesis and testing capabilities to prepare and evaluate new materials. This will involve the fabrication and assembly of our tools as well as the integration onto our IoT platform by developing the control software.

The second is performing the experiments using the high throughput experimentation through a proper design of experiment, to generate high quality experimental data.

The third, will be to manage that data, including the proper visualization, curation and archival of data. The entire experimentation workflow will be driven by machine learning where new experimental conditions will be suggested by a machine learning algorithm that is trained on existing experimental data.