Customer Signal
Interview notes, repeated pains, workarounds, and decision moments that reveal what actually matters.
Product side
Market Atlas
Understand people. Find the truth. Prioritize what matters.
Builder side
Build Engine
Design, build, and ship with AI-native workflows.
I bridge customer reality and technical possibility to build software people love and businesses need.
Explore My Work1M+Daily users reached
0 → 1AI product, shipped
1,500+Early RAG users
6 teamsRoadmap ownership
EpicClinician workflow
Product operator
Customer discovery, prioritization, launch thinking, and decision clarity.
AI-native builder
Claude, Codex, agent workflows, prototype loops, and hands-on shipping.
Healthcare + operations lens
Comfort with messy workflows, constraints, and real-world adoption.
Learning in public
Visible artifacts, build notes, code, experiments, and honest iteration.
Market Atlas
Good products start with customer reality.
The product side is about listening hard enough to separate true demand from noise, then making the tradeoffs that move a team forward.
Problem Framing
Jobs-to-be-done, constraints, user context, and the crisp problem statement behind the roadmap.
Prioritization
Impact, confidence, effort, timing, and the discipline to avoid building what does not matter yet.
Roadmap Bets
Sequenced bets that connect customer value, business context, launch risk, and measurable learning.
Build Engine
AI is leverage for people who know what to build.
The builder side is about using modern agent workflows to turn product judgment into prototypes, shipped pages, useful systems, and better questions.
Plan
Translate the customer problem into scope, acceptance criteria, and a buildable path.
Build
Use Codex, Claude, and code review loops to move from idea to working artifact.
Test
Verify behavior, accessibility, responsiveness, links, and deploy readiness before calling it done.
Ship
Deploy, measure, learn, and feed the next customer signal back into the system.
Selected work
Proof beats polish, but the best work has both.
These cards are the first version of the portfolio model. Each should grow into a concrete artifact with the problem, decision, build, and outcome.
Cardinal Point Services
Positioning and product thinking for small-business services, with the customer problem kept at the center.
Medkit
Offline medical assistant: a Rust core, a local LLM, and a curated corpus, designed around hard constraints and real-world reliability.
This Website
The proof is the page you are on: concept, agent planning, hand-checked code, browser QA, and a manual Netlify deploy. Source is public on GitHub.
Operating loop
Listen, map, decide, build, test, ship, learn.
The loop matters because it keeps the work honest. Product without building can drift. Building without product judgment can sprint in the wrong direction.
- ListenFind the real customer friction.
- MapMake the system visible.
- DecideChoose the highest-leverage bet.
- BuildTurn the bet into a working artifact.
- ShipRelease, measure, and learn.
Contact
Have a messy customer problem or product idea?
I like turning unclear terrain into working software.