May 2015 – Present · 11 yrs · San Francisco Bay Area

• Meta Superintelligence Labs - Building infra for MSL researchers and developers
• AI Infra -
• I support all the Systems, Components and Experiences for AI Training and Inference at Meta. My teams are responsible for enabling teams at Meta to quickly experiment with innovative new modeling ideas through offline, online and recurring training and inference at massive scale and freshness.
• AI Developer Infrastructure -
• My teams support all of the tools and experiences for developing, operating and releasing production ML models at Meta. We are responsible for improving velocity of exploration and experimentation for ML developers, allowing rapid iteration with minimal time building code
• Monitoring & Observability -
• My teams are responsible for several large scale distributed systems -- including our real-time monitoring platforms, analytics tooling, AI/ML investigation tools and visualization products. Some of the work from my teams:
• Fast Analytics using Scuba - https://research.fb.com/publications/scuba-diving-into-data-at-facebook
• High performance Time Series Storage Engine - https://code.facebook.com/posts/952820474848503/beringei-a-high-performance-time-series-storage-engine/
• Multi-dimensional In-memory DBMS https://research.fb.com/cubrick-a-new-multidimensional-in-memory-dbms/