AI operations, enablement, and workflow tooling
I build practical AI operations and workflow tools shaped by enterprise support, Lean operations, and frontline implementation. My public work focuses on cost visibility, evaluation, governance, MCP operations, documentation, and reliable handoffs between people and AI systems.
I am pursuing hands-on mid-level roles in AI implementation, enablement, technical operations, solutions, workflow automation, process engineering, and business systems in Oregon, Washington, or U.S.-remote teams.
| Project | Problem it addresses | What it demonstrates |
|---|---|---|
| mcp-manager | MCP configuration is fragmented across AI development tools | Configuration normalization, health checks, safe write-back, CLI design, documentation |
| ai-spend | AI API costs are difficult to see across providers | API integration, local data handling, cost visibility, practical terminal UX |
| agent-lint | Agent workflows can hide unbounded cost and reliability risks | Evaluation, governance, static analysis, failure prevention |
| arete-evals | LLM behavior needs repeatable evidence rather than impressions | Evaluation practice, rubric design, run records, comparison discipline |
| the-human-stack | AI implementation needs operational controls and reliable handoffs | Process thinking, technical communication, evidence grading, implementation methodology |
- Frontline implementation: process discovery, requirements translation, troubleshooting, standard work, SOPs, training, and adoption.
- Operational improvement: at Toyota Logistics Services, built an internal GPT and low-code workflow connecting scheduling, parts, and production work; combined with root-cause analysis and standard work, this helped reduce downtime 15% and move missing-parts incidents from a 15% baseline to zero.
- Enterprise support: at IBM, achieved greater than 95% first-time tier-two resolution and helped reduce escalations 20%.
- Public engineering practice: Python, SQL, FastAPI, Pydantic, REST APIs, PostgreSQL, SQLite, Docker, GitHub Actions, pytest, Playwright, MCP, and OpenAI and Claude APIs.
- Understand the user, workflow, constraints, and failure modes.
- Translate the work into requirements, controls, and measurable outcomes.
- Build the smallest useful implementation and test the risky paths.
- Document the system, train the people using it, and make handoffs explicit.
- Observe results and improve the workflow using evidence.
My public repositories are independent, production-minded projects. They demonstrate how I build and reason; they are not presented as employer production deployments or enterprise customer implementations.
I am especially interested in work that combines technical problem solving with implementation, support, evaluation, governance, integrations, documentation, training, and adoption. The best fit gives me direct exposure to production AI work while keeping me close to users and daily operations.