Jason Yim

Generative protein modeling.

AI4science

Science program - proteins. Previously DeepMind health research.

Course assistant for 601.482 Deep Learning

Research assistant in the Neurodata lab (neurodata.io/). Research and building tools to analyze brain tissue at the tetrabyte scale. Experience in: - Developing and scaling CLARITY (https://www.nature.com/news/see-through-brains-clarify-connections-1.12768) brain volume registration, neuron tractography, and neuron detection algorithms. Statistics research: - Random Forest (https://jmlr.org/papers/v21/18-664.html) - Linear Optimal-Low rank projection (https://www.researchgate.net/publication/325145804_Linear_Optimal_Low-Rank_Projection)

Instagram Growth team - Designed new database schema and migrated all existing user settings to a new production database. - Designed and deployed Instagram settings page on iOS, Android, and web for email and text notifications as well as add features for promoting monthly user retention. Measured impact with A/B testing and added user growth of ~2k+ monthly active users. - Implemented short ig.me URL for shorter links in notifications. Shorter URLs reduce cost by thousands of dollars by sending shorter SMS notifications. - Prototyped Machine Learning pipeline for promoting user growth through improving user tailored push notifications.

Teaching assistant for 600.463, Algorithms.

Azure cloud team. - Designed and implemented distributed cloud benchmark tool for Azure.

Robotic planning AI research using a declarative framework. Advised by Michael Gelfond.

Co-advised by Regina Barzilay and Tommi Jaakkola