How Engagements Work
From discovery call to delivered evaluation harness. Clear process, fixed pricing, no surprises. Here's exactly what to expect.
Free Discovery Call
A 20-minute call where I listen to what you're building, understand your AI stack, and tell you honestly whether I can help. No pitch, no pressure.
Scope & Proposal
If there's a fit, I send a clear proposal within 48 hours: what I'll evaluate, how long it takes, what you'll get, and what it costs. Fixed price, no surprises.
Deep Dive & Evaluation
I get access to your system, map failure modes, build test scenarios, and run systematic evaluation. Daily async updates via Slack or your preferred tool.
Results & Recommendations
You get a clear report: what works, what breaks, where the risks are, and exactly what to fix. Plus the evaluation harness itself — it's yours to keep and run.
Ongoing Quality (Optional)
For teams that want continuous coverage: monthly eval refreshes, model comparisons, regression monitoring, and release gating. Keeps quality high as your product evolves.
How I Work
Principles that make remote async engagements with US teams actually work.
Async-First
I overlap 4-6 hours daily with US Eastern and Pacific. All communication is async with daily written updates. No meetings unless you want them.
NDA from Day One
I sign NDAs before every engagement. Your data, prompts, and evaluation results stay confidential. Always.
Your Stack, Not Mine
I integrate with your existing codebase — Python, TypeScript, or whatever you're running. No forced rewrites or vendor lock-in.
You Keep Everything
The evaluation harness, test scenarios, scoring rubrics, and CI integration — it's all yours. No dependency on me after the engagement ends.