The term I spent working at IA was extremely eye opening and rewarding. Although I've had experience working at startup companies in the past, working at IA was an entirely different experience with it immediately becoming clear that my contributions were vital to the progress of the company as a whole. As one of two people working on the code base, I was the lead developer on the backend system for a company that was in the midst of rapidly expanding their client base which meant that a lot of older code had to be rewritten and sometimes completely redesigned to allow the company to keep up with their growing data needs. This meant I was given plenty of opportunities to make an impact on the systems I was working on that I knew would continue to matter after my term was over including things like redesigning database structures. I also had the pleasure of working very closely with healthcare professionals to develop new core functionalities and graphs that would allow us to analyze and display our data in new impactful ways to users who have a bit more trouble interpreting information. The work I did at IA was rewarding in a way that I haven't often experienced at other companies where the work I was doing was immediately being used in production with real customers who were relying on IA for their health and these customers were able to return with direct feedback with ways the products could be improved.

In 2020, I had the privilege of working at Uber ATG to develop the technology that would allow autonomous vehicles to safely share our streets. The team I was on developed complex machine learning models which were able to detect and label important things, such as pedestrians and vehicles, using LiDAR data. While I didn't get a chance to directly improve these models, I worked closely with other members of my team to research, propose, and implement light weight metrics to monitor the performance of these models without relying on ground truth data, which is extremely expensive to produce. Since the metrics I implemented didn't add any significant overhead, we were able to collect these metrics in real time in Honeycomb and AWS. In addition, while analyzing the results of these metrics, we discovered that one metric could also be used as a powerful debugging tool for model training. Over the course of the summer, I presented bi-weekly deep dives about my project including a final deep dive and project showcase presented to the entire company.

Despite being funded by a large bank, working at Live Labs really felt more like working at a tiny startup with a small team of 3 full-time employees and 3 other co-op students. Along with a UX designer and a product manager, we created, from ideation to release, 3 pilot projects including native iOS and Android apps leveraging two different types of machine learning models. The projects we worked on were incredibly diverse so I was required to very quickly pickup tons of new languages and frameworks which allowed me to really grow as a software developer. Since we were such a small team, we were expected to act as full-time employees and often ended up becoming the team expert on a pilot. This gave me a really strong sense of ownership over the products that I built and the work I did; so much so that I received a rating of 'Outstanding' on the final work term evaluation for my “incredible ability to learn new software and take ownership of [my] work”.

Thanks to the smaller company size and the flexibility of my team, I had the amazing opportunity to work on four different key products. I gained exposure to a wide variety of programming languages and frameworks including everything from the huge variety of available javascript frameworks to OpenGL to Ant Design and React. In addition, although I initially started out doing the things I was hired for, like manual testing iOS apps and developing basic unit tests, I was quickly able to move on to independently creating complex integration tests, adding new features on a variety of platforms, and even implementing a huge new animation feature on our Android SDK.