I build AI systems that turn expert judgment into reliable workflows.
My work sits at the intersection of business analysis, agent governance, and harness engineering: state machines, hooks, validators, evidence ledgers, and quality gates that make AI work more reliable, reviewable, and useful in real decisions.
- Alpha Insights -- an open-source business research Skill for Claude Code and Codex Desktop. It turns analyst methodology into a 7-stage workflow with 19 frameworks, source confidence grading, quality gates, and polished HTML reports. Demo report
- Alpha Insights-Data -- a strategic analysis system for internal data, extending the Alpha Insights harness into evidence tracking, governed insight generation, and enterprise decision support.
- Anchor -- a Codex skill for agenda-stack tracking, nested TODO control, and drift prevention in long-running AI work.
- Overwatch -- a real-time second-opinion layer for AI agents: session-level oversight, independent review, and architecture feedback while work is happening.
- Forge-PPT -- a presentation generation system for deterministic layouts, image-backed visual assets, and executive-grade decks.
- Convex -- a structured futures and options research system for market scanning, signal review, and disciplined opportunity analysis.
- AI Data Insights Series -- public writing on enterprise AI analytics, governed insight generation, and the shift from prompt-only work to engineered workflows.
- Harness over prompts.
- Evidence over vibes.
- Workflow memory over chat history.
- Quality gates over "looks good."
- Domain expertise should be encoded as systems.
Building reusable AI Skills and agent systems that help analysts, operators, and builders move from one-off chat output to repeatable expert workflows.
Creator of Alpha Insights, Alpha Insights-Data, Anchor, Overwatch, Forge-PPT, and Convex.