Menlo Park, California, United States
Lead for Business Compromise in Cross-Meta Integrity (XI). Worked in detection, investigation, and recovery of compromised Business Manager clients to secure platform integrity for Meta’s 700MM+ advertisers. Developed automation tooling for investigation and recovery, ML pipelines and feature engineering, credit and payment risk frameworks for all Meta Business Manager clients generating $95B+ annual revenue - including all High Value Businesses ($80B ARR) and Small Spend Advertisers ($15B).
Data Engineering and Analytics: Developed and maintained critical data pipelines and frameworks used cross-functionally by Meta Integrity teams to triage, prioritize and recover compromised businesses and for topline metrics and goalin. Owned and maintained primary source-of-truth critical infrastructure datasets and Business/User Graph relationship datasets, deployed ML models and created feature engineering datasets for machine learning detection classifier models.
Data Stack/Tooling: Hive, Spark, Presto/Trino, MySQL, Python, Dataswarm (Airflow), Bento (Juypter), Tableau/Unidash/Daiquery (SQL IDE + Dashboarding tools), Scuba (Grafana equivalent), Loggers, Backfill
Software Engineering: developed Autopilot, a system to automate point-of-compromise detection and Automated Investigation and Recovery (AIR). These were internal tooling and infrastructure launched as central tooling for Business recovery workflows used quarterly on tens of thousands of businesses, including multi-million dollar High Value Business clients. Performed full-stack work with backend focus: primarily product tooling, logging, and integration with Data Warehouse, and some UI development.
Tech Stack/Tooling: Hack (PHP), EntQL (GraphQL wrapper), React/Javascript, chronos Python ETL, Scuba (in-memory realtime analytics DB), Laser (in-memory key-value store).