Staff Technical Program Manager @ Uber, Microsoft

Principal PM @ Microsoft working on Azure for the US Government. I know Cloud Compute, Storage, and Networking pretty well and DevOps (ie: Deployment, Incident Management, and Monitoring) and how to build these things in restricted Cloud environments REALLY well. I also worked about 3Y with Hololens and the MSR team building ML infrastructure for generating synthetic data at scale for computer vision models and K8s based training/inference for ML models on top of Azure. I also worked at Uber as a Staff TPM on their ML Platform team, Michelangelo, and spent about two years driving the adoption and migration of all of Uber's most important ML driven business apps (eg: ETA prediction, Recommendation and Rankings for both Rides and Eats, Automated Promotion Campaigns Pricing) to deep learning models. I also worked at a startup called Anyscale who runs Ray: a distributed machine learning framework that's powering most of the generative AI and LLM space, both on training, data processing, and serving. I worked on two main things; re-architecting the control plane for the commerical offering of managed Ray (lots of infra level stuff and k8s related stuff... lots to do with networking security when building saas solutions for enterprise as well) as something called called Ray Compiled Graphs which is a pythonic acceleration library for optimizing distributed inference and training for large machine learning models. I'm currently working at OpenAI on Applied (ie: our production inference stack) Infrastructure (myriad of things... from our deliver pipeline, cloud based infrastructure, online database systems, our data flywheel and batch processing platform, lots of time on reliability, and a bit on our observability stack).