Software architecture · Applied AI · Product engineering
I work with founders and product teams when architecture, applied AI, or delivery has become hard to reason about. We make the trade-offs visible, choose the next move, and stay close enough to the build to make it work.
Where the work connects
Most hard product problems are not just code problems. The product decision, technical design, delivery plan, and operating reality have to make sense together.
AI that does a real job in the product.
I design agentic workflows, retrieval systems, document intelligence, and evaluation loops around a clear user or operator decision, not around an AI demo.
Cloud architecture the team can understand and change.
I shape service boundaries, data flows, infrastructure, observability, and delivery practices so the product can grow without making every change risky.
Product thinking and engineering in the same room.
I bring product judgment and hands-on engineering together to clarify the problem, choose the right scope, and ship a coherent first version that can evolve.
A decade inside the work
Since 2014, I have worked across early ventures, enterprise systems, insurance, retail, and applied AI. That range helps me see how a technical decision changes the product, the operation, and the team around it.
Independent venture, currently in stealth
Building an independent product practice around applied AI, reliable software, and shorter paths from an unclear problem to a working product.
LuckyTruck
Led engineering for a commercial trucking insurance platform, spanning product delivery, AWS architecture, document intelligence, integrations, and operations tooling.
UVERA
Worked on platform architecture, delivery systems, and service boundaries for a product built around high availability.
Selected systems
Insurance, retail, applied AI, and marketplaces. In each one, the architecture had to make sense to the people selling, operating, or using the product.
Explore all case studies
A connected digital insurance platform designed around the pace, documents, and compliance realities of modern trucking businesses.
A marketplace where customers can compare and buy health, travel, auto, and life insurance, then get help with claims.
A point-of-sale platform connecting transactions, inventory, retail CRM, order management, and real-time stock data.
An in-house OCR workflow that extracts driver-license data, reduces manual entry, and gives operations teams a clear review step.
Where I am most useful
I like problems with real constraints and no obvious answer. Direct questions make the trade-offs visible, and staying close to the code keeps the strategy honest.
How the work gets doneHow the work moves
Clarity before velocitySpeed helps after we can name the real constraint. Until then, moving faster usually means producing more uncertainty.
Start with what is stuck
You do not need a polished brief. Share what the team is building, what keeps getting in the way, and what needs to happen next.