A mix of client work (shown anonymized) and personal builds, grouped by type. Every one is a working system, not a concept. Click any project for the full story.
Three AI "firms" (McKinsey, BCG, and Bain personas, ~5 agents each) compete on one brief, cross-review and rework each other's output, and a partner agent synthesizes the final report. ~26 agents per run.
A team of specialized agents (sales, marketing, ops, account management) run by a conductor that classifies each task, routes it to the right agents in parallel, and synthesizes the result. A small team that operates like a much bigger one.
A set of Claude Code skills I built and use daily. One mines my entire history into a working profile, one writes clean session handovers, and one chains skills together into multi-step workflows. Tooling that makes the AI itself more useful.
Inherited a 4-year-old HubSpot CRM that several leaders had run in and out of. In about 30 days I cleared out roughly 72% of the records (120k contacts to 35k, 60k companies to 15k, 1,200+ deals to 400), then built the hygiene to keep it clean and mined past outbound to sharpen new messaging.
An executive dashboard for a national 3PL that pulls dozens of KPIs from several systems through an ETL pipeline into a Postgres warehouse, so leadership can see the whole operation in one place.
The data backbone under the dashboards. An ETL pipeline that consolidated multiple operational systems and backfilled over a year of history into one clean warehouse the rest of the analytics could trust.
Analyzed dozens of recorded sales calls to find which questions and messaging actually moved deals, then fed that back into the outbound playbook so the team led with what worked.
A 3D model of a ~100,000 sq ft warehouse you can walk through, with light KPIs mapped right onto the floor. Click any zone to see its key signals. The bridge between a dashboard and the physical space it describes.
A role-based internal portal and tool launcher with Google SSO and an admin panel. Onboarding and offboarding run through a staged-diff approval flow, so every change to the live database is reviewed before it applies.
A personal operating system for my week. Recurring calendar structure, a morning automation that surfaces the day's priorities, and a habit tracker built to sync with Postgres and Google.
A full mobile app for lacrosse training, built end to end with React Native and a Node backend plus its own admin panel and a Postgres database. A complete product, not a prototype.
A learning platform that takes non-technical teams from beginner to confident builder: modules, a knowledge hub, a workflow library, and progress tracking. AI adoption made real.
Give it any software name and it researches the docs and generates a full course, both a written guide and a deployed interactive web version. Curriculum work that used to take weeks, done in minutes.