• Storage lead for the AI Research SuperCluster: https://ai.facebook.com/blog/ai-rsc and another datacenter

• Started dataloading library project used across FAIR, Reality Labs, GenAI, MSL (Meta Superintelligence Labs)

• worked in various organizations: FAIR, Storage, MSL Infra

Started at Facebook, switched to FAIR in February 2020

• scale, performance, security, and privacy for AI training

• storage and delivery infrastructure for training content

• distributed coordination, management, and monitoring of performance canary system

I developed ETector, a software plagiarism detector used by a number of leading universities and also licensed commercially.

• sophisticated detection from textual matching through relabeling and reordering

• typically detects and accurately ranks 2x the incidents as other detectors in intro courses

• produces evidence that is easier to comprehend

I managed an experimental systems group working on software for our next-generation hardware stack.

• one system announced at Google Cloud Next (then known as GCP Next)

• other systems deployed internally