Hi! My name is Om Kelkar (she/her). I work on building large-scale, reliable software systems.
Currently, I'm a software engineer working on compute fungibility between large-scale GPU platforms at Meta to promote fleet efficiency. I helped to develop IPNext, an AI/ML Deployment+Serving platform at Meta, turning ML models into queryable web services with support for heterogeneous compute, regional failover, auto-scaling, update safety, and more.
Previously, I worked on distributed tracing and context propagation at Meta and as a full-stack engineer on the Microsoft Viva Insights team. Before all that, I was a student just learning what a program was. Computers are cool.
Outside of the office, I take on long-distance endurance challenges, shoot/edit photo/video, and generally make the most of living in NYC. Bikes are also very cool. Sometimes I write about my trips by bike.