The code is the easy part now. Trust still has to be engineered.

Hi, I'm Bryce, an independent engineer (ex-eBay, nearly a decade). I design and build AI systems for solo builders, startups, and established orgs.

I work with Claude Code , MCP , and whatever else fits the problem, from architecture and quality verification through to production.

See how I work with clients →

Latest week
jul 20 - 26
57
hands-on sessions
17
commits
7
projects

From my weekly work log. See the full log, or the goals it rolls up to.

Recent work

Client engagement

image/svg+xml knowledge-management pilot with Akaya (Oregon AI Accelerator '26 cohort), through Dec 2026.

Open source

MCP API Bridge turns any REST API into an MCP server. Starter kit with 74 tests, source on GitHub.

Upstream bug reports

Traced and filed, acknowledged by Anthropic: #32561 and #32213.

Just published

I had Claude write a RAG eval harness. Here's how I made it auditable.
New

I had Claude write nearly all of a two-layer eval harness for a client's RAG app. The design calls are mine, and they're what make it trustworthy: the judge model's own statement is stored next to the verdict my code derives, so a bad grade contradicts itself somewhere I can see. What I measured, what I rejected on the numbers, and where the numbers still can't help me.

Latest writing

Let's Connect