Could tell us a little about yourself and about your studies before coming to Imperial?
I did my undergraduate degree in Natural Sciences (Physics) at St John’s College, Cambridge, before coming to Imperial. I then worked for two years at Bank of America Merrill Lynch in their Debt Capital Markets division in London.
What attracted you about the MSc in AI?
As someone who didn’t learn to code at a young age, I was very hesitant to make a career move and do a master's in a computer science related area. The specific targeting of the MSc in AI—for those with a quantitative background but no prior coding experience—really gave me the confidence to apply! Also, the modules offered ensured that we got a really good grounding across all areas of AI, rather than simply diving into deep learning. We were taught the underlying mathematics, symbolic AI and reinforcement learning, and were also given the chance to understand the key philosophical and ethical questions in the field. Furthermore, the support available from the DeepMind scholarship for this course was an amazing opportunity and truly enhanced my experience during the MSc.
What did you enjoy the most?
The software engineering group project! Especially for those without a coding background, it was an incredible opportunity to work as a team with an external start-up and learn more about agile development, how to design and build tests for efficient, production level code.
What did you find more challenging?
In the first term, we take machine learning modules in parallel with our Python Programming module where we learn to code. As the machine learning modules also had students who could already code, the coursework was definitely intense. However, our Python module was so thorough and the teaching assistants and instructors were so incredible that we felt supported through the entire process and by the end of the first term were fully up to speed with the other students.
Could you tell us about some of your achievements on the MSc that make you proud?
I was really happy about our results in the Natural Language Processing coursework. In a group of 3, we designed and implemented models to assess the funniness of news headlines. We had a really fun time doing the project together and achieved a great result!
What did you do in your spare time?
I did the MSc in AI in the 2020–2021 academic year, when we were subject to lockdown owing to the pandemic. In spite of that, we made a great group of friends in London and went for walks, bouldering and many brunches and dinners together!
Could you tell us about your individual project?
The title of my individual project was ‘Deep Learning for Stratifying Fibrosis on Histopathological Images of the Liver’. Working closely with an expert pathologist at St Mary’s Paddington hospital, we were the first people to explore a new dataset of images. This was really exciting as it felt like a start to finish project with real meaning. The question was as follows. In clinical trials for liver disease one of the main challenges is the high cost and poor inter-observer agreement when a pathologist assesses fibrosis severity from biopsies. Can we take steps to improve this through deep learning? We developed a robust data processing pipeline and got some pretty encouraging results! Next year some more students can extend the work which is also a really nice feeling.
What have you been doing since you graduated?
I did an off-cycle internship in Artificial Intelligence at JP Morgan, before moving to New York to start my PhD at Columbia University! I'm supervised by Professor Andrew Laine and Dr Elsa Angelini, and in the lab I joined we are currently studying the structure of the lungs using deep learning from CT scans. We collaborate with a huge team of doctors who guide the machine learning research and I’m absolutely loving it!
Do you have any advice for prospective students?
Make the most of the tutorial sessions you are offered! They’re honestly invaluable and also the teaching assistants are so friendly. I asked so many questions that I thought were ‘silly’ or ‘obvious’ and they never judged and explained everything really well!