AI/ML Product Manager at Meta. I share agentic workflows/skills that I use to build my projects - because the best product thinking comes from hands-on building.
I work on Ads recommendation systems at Meta. Previously built AI products at Covariant (robotics foundation models), Google (Pixel on-device ML), and Kami Vision (computer vision).
- AI-native PM workflows — structured multi-agent frameworks that plug into Claude Code and make product managers faster at strategy, metrics, writing, and decision-making
- Knowledge systems — a PARA-based second brain for Claude Code that organizes projects, areas, and research across sessions
- Developer tools — reusable skill architectures for AI coding assistants that compound over time
| Project | What it does |
|---|---|
| pm-copilot | Agent-team powered PM workflows — parallel skeptic panels, cross-pollination debate, and human-gated phases for investigation, strategy, and proposals |
| pm-skills | 17 PM frameworks for AI-assisted product thinking — PRDs, metrics, strategy, prioritization |
| compound-product-management | Claude Code skills that learn from sessions: clear writing, prompt coaching, continuous improvement |
| para-second-brain | PARA knowledge management plugin for Claude Code — structured capture, review, and retrieval |
| claude-skills | Open-source reusable instruction sets that make AI coding assistants smarter |
- Product Manager, AI/ML at Meta — leading supply personalization and ad ranking/quality for Facebook's 3B+ user monetization platform
- M.S. Computer Engineering, Georgia Tech (4.0 GPA) — ML, computer architecture, game AI
- Previously: Covariant (robotics foundation models → acquired by Amazon), Google (Pixel AI, Tensor SoC), Kami Vision (CV-powered consumer products, $36M ARR)
- Built 27+ Claude Code skills and AI-powered productivity tools as daily-driver workflows
I build AI products that ship to real users. Research capabilities only matter when they become products people trust and use daily.