Meet Summation Apprentice Jane

21-year-old Jane Seah is a fourth-year Information Systems student at Singapore Management University, who had recently ended her apprenticeship with INTELLLEX. For four months, she helped to build an automated end-to-end system to extract useful information from legal documents.

Read on to find out Jane’s thoughts on the Summation programme and her INTELLLEX’s experience.

After my ‘A’ levels, I searched for possible courses to take in University and programming was one of the top skills to have in the near future. I started learning to programme on my own and grew to like it. I also saw the need to understand business problems to apply appropriate IT solutions, so I decided to enrol in Information Systems.

I worked on the ConanD team, which is a combination of legal knowledge engineers and data scientists. The legal knowledge engineers analyse the search patterns of lawyers and try to understand the results they desire. With the help of their legal subject expertise, we develop machine learning capabilities that are refined to lawyers’ needs.

I worked on three main arcs during my apprenticeship:

1. Implementing features and fixing bugs
2. Testing out the classifiers
3. Project JANE — legal word embedding for the machine learning systems to deeper understand the context of documents. (The project name was not my choice of name!)

The nature of machine learning is that the training of models takes significant time. However, I was never sitting around waiting. There were always bugs to fix or small tasks to do. This was a good taste of what working life is like because you will often be juggling multiple projects.

My significant achievements were:

1. Learning a new programming language and paradigm, successfully implementing features and solving bugs on the product which real clients use
2. Spearheading a research project and developing a prototype to amplify search hits on INTELLLEX products further; delivering an internal use web service to illustrate the effectiveness of the solution.

First of all, I think I was really lucky because usually with internships, you don’t know what you’re getting into. With INTELLLEX, I got to build and train models and work on the actual product that lawyers and clients are using. This gave me a lot of validation because they trusted me to have code in their actual product.

It was great to learn from my supervisor, Archibald, who has applied experience and extensive knowledge about machine learning. He could foresee the usual problems and mistakes I might make and would give me advice and direction. He was also supportive. I had some self-doubt, but he gave me a lot of confidence by always giving me due credit.

I also had the opportunity to interact with different people across functions regularly. I also really liked the company huddles every two weeks, because I got to hear from management where they shared their business vision and strategy, as well as future product direction. From there, it was much easier to link my work with the company’s direction so I could see how I was tangibly helping to make the product better.

I think one of the challenges I faced was learning Clojure, which was a completely new programming language for me. It was quite challenging because I was told that by the end of my first week of learning, I should be able to write a feature in Clojure. I didn’t think I could, but I eventually got the hang of it.

Another challenge was working on Project JANE – I would have loved to work on it more. During my internship, I felt that I didn’t have enough knowledge about word embedding models. I realised that one of my biggest obstacles was not having strong math fundamentals, so I am going to take up more math courses.

What motivated you to join Summation programme?

Having interned at a large government agency previously, I wanted to experience what it was like working at a growing startup. It was also important to me that "AI for good" is at the core of the company I work at. Furthermore, it is not easy to find a data science internship where you actually get to do data science and work closely with the business and the software engineering teams.

I also took the opportunity to learn more by attending two workshops organised by SGInnovate – Deep Learning Jumpstart and Advanced Natural Language Processing. The workshops helped to strengthen my grasp of the concepts behind the different Deep Learning models. As a result, I have a better sense of which models work better for which problems and have an enhanced intuition for tuning Deep Learning models.

Our Summation programme enables apprentices to work in technology-intensive companies on Deep Tech projects such as AI, machine learning, and deep learning. Be a part of this journey here!

This article was first published here.