Open to remote senior and staff PM roles
Staff PM for pricing, payments and point-of-sale lending. I use AI to turn real problems into working products.
8+ years building the pricing, credit and checkout systems that sell and finance home solar, at Palmetto, ADT and Patter AI. Today I use AI to go from problem to working product in days, starting with problems in my own life and putting each one in front of real users.
Career highlights
From first PM to staff, in pricing, payments and point-of-sale lending.
Patter AI
An AI-native sales platform for solar and home services. I led a single real-time pricing engine across loans, leases and PPAs, a rep-guided checkout that shows the payment impact of each add-on, and a lender rules engine that cut onboarding from about a month to a week.
ADT
First PM for a 3,000-person solar division. I built the credit-check and offer engine (12,000+ applications a month across 5 lenders) and lender allocation that helped cut dealer fees 3–5 points, with releases to about 5,000 reps.
Palmetto
I was Palmetto's first PM and built Alchemy, its design, point-of-sale and real-time credit platform, and supported a $6M venture round. Then I launched a machine-learning lead marketplace at 11,500+ consultations a month.
What I've built with AI
Each product started with a problem in my own life. Together they show how my practice matured, from a single prototype page to a team of agents working under tests and review. Every demo runs on sample data in your browser.
Energy Scanner
Bedtime with my kids had become a nightly negotiation. I designed a routine that ends in an energy scan, so a neutral signal, not a parent, calls lights out.
Built as a single page in one evening and tested with its real users at bedtime that night.
Role Radar
A job search is mostly triage. An agent now reads about 110 company boards each weekday and scores every role with a stated reason, so my time goes to the few worth pursuing.
My first scheduled agent: it runs unattended, and I review its judgment rather than the raw feed.
ccwatts
I wanted to know the electricity cost of my own AI-assisted coding. No tool answered it, so I built a local one that reports a sourced range instead of a false point estimate.
My first open-source release, with a public methodology, full tests and a citation for every assumption.
Family photo app
Two blended families, five group texts and grandparents on Android. I'm building one private place for our photos, where every post reaches exactly the people it's meant for.
I ran AI agents as a product team: a spec, tests first, a privacy review and a usability pass for grandparents.
Family beta with about a dozen relatives, mostly grandparents, in late November
How I keep it all moving
Running five products in limited hours takes a system: one slate, one roadmap and AI agents with clear ownership.
The operating layer under everything else. The demos run on a scrubbed snapshot of my real slate.
The studio behind it
Five products and one person. Each AI chat lost what the others knew, which made me the only shared memory and the bottleneck.
Specialist agents, a coordinator that only coordinates, and me as product owner, reviewing what ships.
Every product on this page runs through it