Member of Program Staff @ OpenAI

I work on the infrastructure used to serve ChatGPT (chatgpt.com and APIs for developers) to millions of users on the Applied engineering team. Previously I worked at Anyscale, The company behind Ray, a distributed computing framework in Python. I spent a year there moving all of our customers on Anyscale, the commercial managed offering of Ray, to a new runtime and another year working on Ray Compiled Graphs, an optimization of Ray for GPU-GPU communication. Before that I was at Michelangelo, Uber's AI/ML platform for deep learning, development experience, and feature embeddings. The platform powered the ML solving a variety of Uber's challenges ranging from safety, to ETA predictions, to Eats ranking and more. I'm pretty familiar the ML pipeline from storage, data, feature prep, transform, training, evaluation, and serving (albeit my bread and butter is still at the infra level... read on...) I also worked with the Hololens, Azure, and Microsoft Research teams in NLP and CV ML (eg: object detection, classification, tracking, search, semantics, and correlations to name a few), specifically around generating synthetic datasets leveraging the power of distributed computing on the Cloud. About 75% of my time working on the above have been on the infra side of things; although I've driven ML performance programs working directly with researchers, data scientists, and machine learning engineers as well! I also spent years working on Azure Government clouds, which are physical, logical, and network isolated instances of Azure for the US government at various levels of classification (from Civilian all the way to S/TS). I know how to build Clouds, and the DevOps required to run them, quite well. In my college years I "founded" two startups (and am always tinkering on something. My (current) passion lies in AI and ML ^_^.