Work with me

Short engagements for AI features stuck between a demo and production.

The two things people pay for

An hour or two. You bring a repo, a failing agent loop, a vague Upwork scope, or a decision you’re circling. I look at the real failure mode instead of the symptom. You leave with code, notes, or a clear next move.
Book on Codementor →

A two-to-six week engagement. Same starting point, but I stay in the repo: ship the PRs, wire the evals, set up the guardrails, write the runbook your team can own once I leave. Engagements are fixed-fee, scope-locked. No retainer creep.
Email me →

What I’m useful for

• Debugging an LLM integration or agent loop.
• Reviewing an AI architecture before you build it wrong.
• Diagnosing RAG, tool-calling, prompt, or context problems.
• Setting up practical AI coding workflows for your team.
• Fixing SwiftUI / iOS architecture or build-time issues.
• Untangling a repo, CI, modularization, or DX bottleneck.
• Deciding whether a prototype is worth productionizing.

What I’m not useful for

• Writing your prompts.
• Picking your model.
• Building a chatbot wrapper around an API.
• Anything that doesn’t have a real user on the other end.

Track record

WhisperBoard 4.8★ on the App Store. mcp-browser-use ~900 stars. Codementor 5.0 across 30 sessions. Five years at Uber on server-driven UI then payments. In-house now at a large enterprise SaaS company running agents in production.

Email: [email protected]