Trading, Research, Data, LLM infra

Ads ML Infrastructure

• Led the cross team effort to redesign Ads Trainer Service with layering

• Metrics Monitoring: designed system to collect Ads Training stack metrics for performance optimization

• Testing Framework: built testing framework for offline training, improving testing efficiency by 20x.

• Ads Core model onboarding, data pipeline and E2E training enablement

• Platform reliability and performance optimization

Proposed and designed predictive models for fraud detection and bot detection

Designed and implemented new Ad telemetry methods

Implemented robust distributed applications to perform streaming classification on billions of data points per day

Working with translational scientists to prototype and productionalize machine learning models for food recommendation and health scores.