Interview with Anna-Lena Popkes, AI Resident
Anna-Lena Popkes, AI Resident
Tell us a bit about your background.
I got interested in AI quite late. I started studying cognitive science in my undergraduate degree and planned to focus on neuroscience and research in this field. Only in my last semester, I came across a machine learning lecture. Afterward, there was no way back. I became so fascinated by the idea of algorithms that can learn on their own that I decided to switch fields, although I had no previous experience with AI or computer science. So I started studying computer science in my graduate degree with a focus on machine learning. It wasn’t the most straightforward path but was definitely worth it! I love what I’m doing. Although I still have many things to learn that other students with a computer science or math background may already be familiar with, I believe that an interdisciplinary background can also be beneficial.
After finishing my graduate degree in Germany at the beginning of 2018, I went abroad for a deep learning internship at the Bosch Center for Artificial Intelligence in Palo Alto in the U.S. My time at Silicon Valley convinced me even more that I chose the right path.
What project are you currently working on?
I’m working on a project applying Bayesian Neural Networks in health care. I was fortunate to be assigned to this project as it was my top choice. I had never looked at the field of Bayesian Learning before, so my AI resident partner and I had to do a lot of reading in the beginning. Now I like the project so much that I could continue working on it for the rest of the residency! Of course, the application in health care makes the project even more exciting and motivates us to achieve good results. In case the project succeeds, its results can be deployed in practice and have an actual impact on the wellbeing of patients.
Tell us about your team environment. What do you enjoy about collaborating with your small team and the larger MSR lab?
At the beginning of the residency, each resident was assigned to a specific project, together with another AI resident. We got matched up such that we have different skills and can learn from each other. That was great! My partner has been incredibly helpful – I don’t know what I would do without her. In addition to an AI resident partner, our project has an engineering supervisor and a machine learning/research supervisor. We also have two other researchers on our team who have been working on Bayesian Neural Networks for years. We meet with the entire team once a week to discuss our current progress, results, etc. We also have lunch together as a team once a week.
Apart from the project team, there is a lot of collaboration among the residents and MSRC in general. The residents help each other whenever possible. Other researchers and engineers in the lab are also happy to help when their expertise is needed, ensuring we get the best possible support.
How have you been designing your ongoing educational opportunities while in the residency?
Many lectures and other events are always happening at Microsoft Research Cambridge providing plenty of opportunities to learn. We also have an AI residency lecture about once a week. These lectures are intended to teach us essential skills like probability theory, etc. We are also allowed to spend some time during the week on learning new things. However, this is a skill I’m still cultivating. I can get so absorbed by our project that I find it difficult to stop working and look at a different topic.
Walk us through a day in the life of a resident.
During my day at work, I am usually coding or discussing new ideas or problems with my AI resident partner. We sit next to each other so it’s easy to talk in case a problem occurs. We often split tasks and review each others code afterwards. Around noon most of the residents have lunch together and then congregate around the ping-pong table. During the afternoon I sometimes attend talks or lectures I am interested in or one of the gym classes. When stuck on a coding challenge this can be quite helpful! Every two weeks we have a one-on-one meeting with the AI residency coordinator. During this time we can talk about anything that we feel is important, including problems or worries. I must say that the support system within Microsoft has been fantastic. There is always someone around to talk or help, no matter what the issue might be.