Product Engineer · Boulder, CO
I build AI software that survives production.
0-to-1 for consumer and wellness brands, plus the internal systems they run on. I work across product, design, and engineering: figuring out what is actually worth building, then shipping it to a real production bar.
What I do
Most AI work looks great in a demo and falls apart in production. I build the other kind.
I am a product engineer. I work across product, design, and engineering, figuring out what is actually worth building, and building it. The result is integrated software a business can trust, from the dashboards it runs on to the full products behind them.
I harden software to an extent that was not practical a couple of years ago: simulating thousands of users across every surface, each pressing every button under every condition, and tracing each piece of data as it moves through the system. The payoff is software that holds up in production, where every data path stays visible and failures surface in testing instead of in front of customers.
Selected work
CertREV
Founder and Founding Engineer
I built CertREV from zero to revenue in three months as the sole engineer: a SaaS platform that certifies AI-era brand content by matching it with credentialed, domain-specific experts.
Architected end to end in TypeScript (Next.js, React, Hono typed-RPC, Drizzle ORM on Neon Postgres, Better Auth, Vercel): a from-scratch Anthropic agent loop, a sandboxed microVM runner for durable long-running jobs, capability-based RBAC with per-brand isolation, and an MCP server that exposes brand knowledge to external LLMs without leaking source data. A 90-table schema, CI schema-drift gates, HMAC-signed APIs, and roughly 190 automated tests.
certrev.com →Custom, per-brand systems for content, marketing operations, and the internal tooling these brands run on:
- e.l.f. Beauty
- BODi
- Intellipure
How I work
- A real production bar
- Staging, automated tests, schema-drift detection in CI, and fail-closed compliance gates. The rigor is what makes the output trustworthy.
- Fast, and on the hook
- AI agents are how I move fast. I am on the hook for every line that ships.
- Built LLM-native
- I build with large language models in mind, so the codebase stays readable and extensible rather than a black box. What I hand off, your own agents can read and extend.
- Clean handoff
- Remote delivery, embedded with founders and teams, with a self-sufficient handoff so the software stays production-grade long after the engagement ends.
About
I studied philosophy at Kenyon College, grounded in epistemology and trust, and I have spent the years since applying that to credibility in the AI era.
I taught myself to engineer and now build production software end to end. The throughline is trust: how you make AI-era output something a business, a search engine, and a customer can actually believe. That question is what CertREV exists to answer.