Ana Sofia Uzsoy is a PhD student at Harvard. Here she talks about her experiences on the 2021-2022 MPhil MLMI.
What was your background before starting the MPhil?
I graduated from North Carolina State University in the USA with undergraduate degrees in both physics and computer science. I was also lucky enough to be able to do machine learning-related internships at NASA and Google. I knew that I wanted to do a PhD in astrophysics, but I was also quite interested in machine learning and wanted to learn more.
What did you get out of the course?
I learned that there is so much more to machine learning than just neural networks: there's Gaussian Processes, Hidden Markov Models, variational methods, and so much more! The coursework in Bayesian statistics and probabilistic modelling was especially interesting and relevant for me, and I continue to use these and other methods I learned throughout the course in my current work. I really appreciated that the course is a mix of coursework and research. I feel like I learned a lot and also gained valuable research experience going into my PhD. As an international student, the UK education system was new to me (and quite different from what I was used to!) but I had a great experience and came out of it with a new perspective.
What were the highlights?
The main highlight of the course for me was really my cohort. It was amazing to meet so many enthusiastic, interesting, and driven people from all over the world with a wide variety of backgrounds and future goals. Throughout the course, we spent many afternoons in the "MLMI room" in the Engineering department working together. Additionally, I particularly enjoyed the Advanced Machine Learning module, where we did a group project extending the results from a recent paper.
What have you gone on to do after the course and what are your longer-term plans?
I am currently a second year PhD student in astrophysics at Harvard, where I use statistical and machine learning techniques to analyse large astronomical datasets. I'm not completely sure what I will do after my PhD, but I will likely pursue either an academic position where I can continue this kind of interdisciplinary work, or an industry position in machine learning or data science.
What advice would you give to people applying for the course?
I would recommend anyone who is considering applying to definitely go for it, even if working in theoretical machine learning is not your ultimate career goal. The skills you would gain are applicable to a wide variety of fields! I would also suggest making sure your background in math/statistics is sufficient, as I found that the ratio of math/statistics to coding in the course was larger than I had anticipated. That being said, I would also consider other similar courses, such as Advanced Computer Science or Data-Intensive Science, and make sure that MLMI is the one that best fits your interests.
What advice would you give to people taking the course?
I would encourage people taking the course to really get to know their cohort and make connections with faculty in the department. I would also advise them to think about their future plans and what they are looking to get out of the course, so that they can choose their modules accordingly and identify faculty with similar research interests. Also, don't stress about your marks: just focus on learning as much as you can. Aside from that, I would just say to enjoy your time at Cambridge; it really is a magical place!
27th February 2024.