CiaraMaria
- senior product engineer
- ai harness builder
- technical leader
Colorado, USA
I build production software and the AI harnesses that help teams ship them.
afk in the garden ✦ humans stay in the loop ✦ last seen: arguing with a type error ✦ coffee count: classified ✦ afk in the garden ✦ humans stay in the loop ✦ last seen: arguing with a type error ✦ coffee count: classified ✦
The short version.
I’m a senior product engineer and technical leader who helps teams move from idea to working software.
- internet since:
- myspace era
- currently learning:
- native iOS because the web wasn’t enough
- probably thinking about:
- another coffee
- creative outlet:
- crochet, gardening, content creation for my dog's insta
Selected work.

AI Harness Engineering
AI Systems / Engineering

Every Hour
Faith / Product Engineering

Sunra Health
Health / Product Engineering

Ryan W. Parrish
Clinical Psychology / Product Engineering

Trellis
Lifestyle / Product Engineering

Closetcore
Lifestyle / Product Engineering

Papaya
web3 / Co-Founder

Stacks + Clarity Camp
Developer Ecosystem / Resident
Where I add the most leverage.
Product Engineering
Taking products from ambiguous requirements through architecture, implementation, QA, deployment, and iteration.
Frontend Architecture
Building scalable TypeScript applications, complex workflows, reusable component systems, responsive interfaces, and maintainable frontend foundations.
AI-Native Delivery
Designing reusable agent harnesses with profiles, policies, fail-closed contracts, guardrails, run-plan schemas, observability, validation, and human review gates.
Technical Enablement
Helping engineers, clients, and nontechnical teams understand systems, make better decisions, and become more independently capable.
AI-native modernization needs more than an agent.
I help teams install domain-specific harnesses: repeatable workflows, policies, contracts, readiness gates, templates, and evidence requirements for real product and engineering work.
Good agent work needs operating boundaries.
- Start with the real user problem.
- Understand the system before changing it.
- Turn ambiguity into a written plan.
- Give agents clear contracts and boundaries.
- Keep humans in the approval loop.
- Verify the output before trusting the speed.
- Feed failures back into the system.
- 1Research
- 2Plan
- 3Implement
- 4Verify
- TypeScript
- React
- Next.js
- Playwright
- Zod
- Figma
- MCP
- Codex
- Claude
- Cursor
- Linear
- GitHub
