AI Engineer
Paid delivery: Acuity Real Estate (January to March 2026), bilingual SMS qualification, 500+ inbound leads. Public proof: DocExtract at 95.5% weighted field-level accuracy on a 28-case offline replay; llm-reviewer-path (clone and pytest, no API key); mcp-server-toolkit on GitHub (617 collected tests).
The reviewer path begins with paid Acuity delivery, then DocExtract evaluation evidence, a no-key clone-and-pytest path, and supporting MCP backend work on GitHub.
Start With Flagship Repos →Strongest technical areas are Python and FastAPI services, retrieval, structured outputs, tool use, evaluation fixtures, and human handoff.
View public proof →Golden evals, adversarial cases, CI gates, prompt versioning, and an 80% DocExtract CI coverage gate make the AI work easier to inspect.
DocExtract on GitHub →Bilingual SMS qualification for 500+ inbound leads, delivered January to March 2026 with CRM integration and handoff workflows.
GitHub →95.5% weighted field-level accuracy on a 28-case offline replay, with a 202-case corpus (151 golden, 51 adversarial) and an 80% CI coverage gate.
GitHub →Clone, install, and run pytest with no API key. Review eval gates, action boundaries, retrieval failures, and scoping receipts.
Python MCP server framework with 617 collected tests, a 30-case adversarial corpus, and support for Python 3.10 to 3.12.
AI Engineer, production LLM systems, agents, and evals. Paid Acuity delivery first, then public proof on GitHub (DocExtract, llm-reviewer-path, mcp-server-toolkit). Contract-first, remote, US-based.
Cathedral City, CA · Remote nationwide · Contract-first; also open to remote AI Engineer roles