• Led the development of Compliance Dynamo, BRM’s first AI-driven compliance research agent. Integrated web scraping, external data gathering, and agent reasoning to automatically populate procurement forms with transparent explanations and citations.
• Delivered measurable results: reduced manual vendor research time from ~1 hour to ~15 minutes for the primary adopter.
• Architected a robust web-scraping system using ScrapingBee and custom JS automation (querySelectors, SPA hydration, scroll orchestration) to provide high-quality context for AI workflows.
• Pioneered BRM’s adoption of tool_call–based AI Agents, working closely with founders to integrate agent tooling into both internal processes and production features.
• Implemented BRM’s first evaluation scripts, introducing repeatable testing for stateless AI functions and establishing a foundation for model quality improvements.
• Accelerated developer productivity by updating our build pipeline, reducing build times by ~80 seconds (90s → 10s)