From theoretical condensed matter Physics, to working with deep neural networks on his ASI Fellowship project with babylon health, Meet Sam, one of our Fellows from our January 2016 Fellowship, as he reflects on his experiences on taking part in the ASI programme.
Tell us about your background, before you decided to move into Data Science?
I did my undergrad in natural sciences at Cambridge, and stayed on for a PhD in theoretical condensed matter Physics. My research focused on building simple "phenomenological" models to describe how currents are generated inside organic solar cells. These exciting devices, formed from the same materials used to make flexible LED displays, have the potential to dramatically reduce the cost of solar power; but they are also poorly understood and difficult to optimise in practice.
I greatly enjoyed this experience, but as I entered my final year I realised I was reluctant to enter the "postdoc lottery", risking years of my life on the hope of a permanent position which might never materialise. As I began to explore alternative career paths I had three main criteria: I wanted a career which would require me to be creative and solve challenging problems, I wanted to write code, and I wanted the security and stability which is lacking in academia. Very quickly, I worked out that Data Science was the natural choice! I began taking online courses in computer science and machine learning, and applied to the ASI.
Why did you chose the ASI Fellowship?
I enjoy learning and solving interesting problems. I don't enjoy filling out job applications, and I also didn't really know who I should apply to anyway. The ASI fellowship provides a way for academics to find out about small Data Science companies that they wouldn't apply to otherwise, and it helps you get a job offer by doing a challenging project and developing your skills, rather than by spending the same amount of time filling out application forms. I also really appreciated the chance to try out a new career path, without really committing myself to anything.
What really surprised me about the ASI was how engaged the company seemed in helping you as an individual, beyond the point at which it is really in their self-interest to do so. Probably the most useful part of the fellowship is not the organised teaching, but the informal conversations I had with the ASI team when I was stuck or needed advice.
What was the best part of your project or experience on the ASI Data Science Programme?
During my project, I worked with a med-tech company called babylon health working on a mixture of deep learning and NLP. To my surprise, I actually learnt more in 8 weeks on this project than I did in any 8 week period of my PhD. I really enjoyed learning about recent developments in machine learning one week, and then applying them myself the next.
The best part of the ASI program itself was the first week of the fellowship, during which all the project companies come in and give a half hour presentation on their project, followed by an extended questions session. This gave me a much better idea of the range of projects Data Scientists work on, and what kind of companies interest me. I'd never heard of most of the companies before their presentations.
What would be your advice to people looking to enter Data Science in Industry?
Brush up on your Python (there are great resources on Udacity/Edx/CodeAcademy), and then take Andrew Ng's Coursera course on Machine Learning. It's written for a wide audience and pretty straightforward, but provides a brilliant introduction to all the key ideas and practical issues you need to worry about. Once you've completed this course, if you feel the need for a more rigorous introduction, google "David Mackay".
It's also worth noting Data Science is a very broad term, and that any two Data scientists might do very different work and have different skills. You'll be able to prepare better if you think a little about what kind of Data Scientist you want to be. Do you want to work at a startup or an established company? Do you want to be a machine learning expert, a generalist, or a more tech-focused database engineer? Would you like to build a product or to have more of a consultancy role? Do you want to learn natural language processing or avoid it like the plague? One thing the ASI really helped me do is answer these questions.
How was your overall experience on the Fellowship, and what have you been up to since completing the Fellowship?
I had a great time on the fellowship. I learned a lot, confirmed that I'd like to work in Data Science/Machine Learning following my PhD, and also really enjoyed getting to know 16 other fellows all much more knowledgable and better prepared than myself! I've already been back to the ASI office once, and am looking forward to the next meetup.
I'm now back in Cambridge for three months, finishing off one last research project and writing up my thesis. The week after my fellowship ended, I went to present my ASI Fellowship project to babylon health where my eight week industry project was based, I am hoping to collaborate with them to write a paper on my project work in the future and discuss my next steps in my Data Science career!