I build the systems between the tools a business already uses: reliable data pipelines, AI agents, and the infrastructure that keeps both running.
I am a cofounder of WhiteCap Data, where I help small and midsize businesses connect fragmented systems and replace manual handoffs with dependable automation.
My work sits at the intersection of:
- Data integration: CRM, billing, communications, spreadsheets, and operational systems connected around a trustworthy source of truth.
- AI and agent engineering: MCP tooling, local and hosted LLM workflows, voice agents, evaluation, and production automation.
- Infrastructure: Linux, Docker, Kubernetes, Cloudflare, private networking, monitoring, backups, and incident recovery.
I prefer practical systems that are observable, recoverable, and useful to the people operating them.
| Project | What it does | Install |
|---|---|---|
| MCPGauge | Deterministic, offline quality grading for MCP servers. | pip install mcpgauge |
| Ollama Handoff | Sends suitable agent tasks to a local Ollama model to reduce cloud cost. | pip install ollama-handoff |
| TableBridge | Sandboxed SQL queries over CSV, Parquet and JSON files without modifying them. | pip install tablebridge |
| WhiteCapData Dev | MCP tools for operating Kubernetes and k3s environments with safe defaults. | pip install whitecapdata-dev |
| Finance Tracker | Local subscription tracking, budgets, alerts and overdraft forecasts. | Run locally |
The four Python tools are developed in public and published on PyPI: MCPGauge, Ollama Handoff, TableBridge and WhiteCapData Dev.
I contribute code fixes, regression tests, documentation and practical troubleshooting answers to community projects.
Verified snapshot: October 8, 2026. These totals cover tracked public community contributions, excluding PRs in my own repositories and internal company work.
| Merged upstream PRs | Open upstream PRs | Discussions contributed to | Accepted answers |
|---|---|---|---|
| 22 | 32 | 45 | 5 |
Browse my upstream PR history.
| Project | Contribution |
|---|---|
| DuckDB | Documented PREPARE, EXECUTE and DEALLOCATE, clarified DECIMAL defaults, and corrected Postgres extension, transaction and mlpack documentation. |
| MDN | Clarified image load timing and session storage behavior, removed an obsolete Android note, and corrected CSS example spacing. |
| MemPalace | Added a regression test for closing a metadata reader before fallback. |
| TRACE | Made fixture generation portable on Windows and repaired attestation reference links. |
| ESLint React | Added detection of mutations through for...of iterators. |
| Cloudflare | Clarified AI Gateway authorization headers for customer API keys. |
| mloda | Corrected input configuration examples. |
| OXVG | Fixed escaping of style content. |
| Sotto | Added CLI share limit validation tests. |
| Pipecat | Repaired a documentation link and corrected wording in Voice UI Kit and examples. |
| FastMCP | Repaired authentication and provider documentation links. |
| MCP server directory | Added Ollama Handoff to the community directory. |
These PRs are open and subject to upstream review.
| Project | Proposed change |
|---|---|
| Pipecat | Resume deferred tool inference when user speech ends. |
| Renovate | Honor Poetry primary package sources. |
| ESLint Stylistic | Fix switch body spacing, JSDoc formatting and nested call spacing. |
| Sphinx | Correct return description indentation and filtered warning links. |
| Ansible | Improve Galaxy service account authentication guidance and Python compatibility documentation. |
| SOFA and NetBox | Fix installed GLFW include paths and document custom script API changes. |
| MCP Apps and Zebar | Improve MCP Apps initialization guidance and Zebar configuration schema generation. |
- NetBox: DNS save behavior and tag ordering.
- Open Notebook: Troubleshooting slow Ollama source chat.
- Wiki.js: Git storage ignore behavior.
- DuckDB: Postgres connection pools and thread contexts.
- FastMCP: Shared HTTPX clients and cookie isolation.
Languages: Python, TypeScript, JavaScript, SQL, Go
AI and data: MCP, Claude, Ollama, DuckDB, Pipecat, agent evaluation
Infrastructure: Linux, Docker, Kubernetes/k3s, Cloudflare, Supabase, Tailscale, OPNsense
I am open to data integration projects, AI agent development, workflow automation and practical infrastructure work for small businesses.
Michael Tierney · Finger Lakes, New York