Senior Product Leader | Human-Centered AI Systems | Pursuing PhD in HCI (Fall 2026)
Website | LinkedIn | Read my AI Philosophy
We are currently in a "Velocity-First" era of AI development that prioritizes speed over safety, often resulting in dissolved context and accuracy collapse. My focus is on building for Integrity-First. If a system is viable and feasible but ethically compromised, it is a failed product.
I build AI systems that enhance human capability and organizational effectiveness, bridging the gap between deep technical execution and operational reality.
- Ember: Authentic Voice at Scale: A multi-agent system currently in private alpha, built to scale authentic human voice without generating "AI slop." It utilizes a custom OWA Architecture (Orchestrator, Worker, Antagonist) to enforce quality and prevent hallucination through rigorous, adversarial review.
- Agentic AI Research: Open-source exploration of Claude code subagents, tools, and Model Context Protocol (MCP) server integrations.
- SpawnForge: An AI-native 3D game engine for the browser. Leveraging WebGPU, Rust, and MCP commands to allow LLMs and agents to build complete interactive environments via natural language.
- Another Way: Why the "Velocity-First" approach to AI development is breaking, and why Operational Integrity is the ultimate competitive moat for modern enterprises.
- The Antagonist Agent: Why traditional AI coding assistants fail due to dissolved context, and why formalizing adversarial roles—an AI designed never to provide positive reinforcement—creates a necessary forcing function for quality.
- AI & Agentic Dev: Claude Agent SDK, LangChain, MCP (Model Context Protocol), Google ADK, Multi-Agent Orchestration.
- Engineering: Python, TypeScript, C#, SQL, Rust.
- Infrastructure: GCP, Azure, Supabase, Vercel, Railway.
- Observability & Analytics: Tableau, Salesforce, Sentry, Posthog, Pendo, Hotjar, LaunchDarkly.
"The biggest leap in AI isn't a bigger model—it's smarter placement of it."