Skip to main content

Projects

Things I've built. Training products, internal tools, and whatever I dive headfirst into next.

Company Work

TrainerRoad

2020–Present

TrainerRoad AI

Took TrainerRoad from reactive to predictive. Simulation engine that runs hundreds of scenarios to find the right workout. 91,000 athletes, 83% of subscribers, used an AI feature last year.

AI/MLProduct LeadershipStrategy

TrainerRoad

2026–Present

Pulse

An internal AI layer that connects Slack, Zoom transcripts, GitHub, Zendesk, and our forum into one conversational tool. Ask it where a project stands. It runs daily digests, links bugs to the code that broke them, and surfaces work falling off track. Available in the terminal via Claude Code or as a Slack app.

OperationsAIInternal ToolsSlack

TrainerRoad

2021

Adaptive Training

Built on Plan Builder. Grades how each workout went and adjusts the rest of the plan. I ran the beta from 100 athletes to 20,000 before it went on for everyone. It now adjusts 72% of plan workouts before they're ridden.

Machine LearningProductData

TrainerRoad

2019

Plan Builder

Writes each athlete a custom training plan toward their race from the time they have and the events they care about. 158,000 athletes have trained on one. Three in four races athletes target on TrainerRoad have a Plan Builder plan behind them. Adaptive Training and everything after it were built on top of this.

ProductPlanningData

TrainerRoad

2025

TrainerRoad x Zwift

Shipped the TrainerRoad-Zwift integration so athletes could do structured TrainerRoad workouts inside Zwift. Two months, small team, direct partnership with Zwift. Athletes who connect do 23% more TrainerRoad workouts.

IntegrationPartnershipProduct Leadership

TrainerRoad

2025–Present

TrainerRoad Accounting Hub

Live financial intelligence for TrainerRoad. Pulls directly from the QuickBooks Online API to give the team real-time projections, P&L analysis, outlier detection, and growth scenarios. Killed the static quarterly forecast spreadsheet and replaced it with a model that's always current.

Financial ModelingQuickBooks APIOperationsAutomation

Personal Projects

Personal Project

2026–Present

NorthStar

I wanted a simple way to track my investment projections over time and see if I was on track for retirement. So I built one. Monte Carlo simulation, CFP-grade tax modeling (federal + state, TCJA sunset, ACA, IRMAA), Social Security joint claim optimizer, and stress tests for real-life events like job loss, long-term care, and losing a spouse.

SwiftUIiOSMonte CarloFinancial Modeling

Personal Project

2026–Present

Tripl

I had a crappy Google Sheet for tracking HSA expenses. Then I opened Claude Code and had a working product in an hour. That was my oh-shit moment. Tripl is now live on the App Store with AI receipt parsing, smart lump-sum reimbursement, and a native iOS app.

Next.jsSwiftUIAISupabaseFull Stack
Visit live site →