Founder, product leader, and builder with a proven track record of turning ideas into working products and real revenue. AI is part of the toolkit — not the story.
With over 20 years building products across SaaS, digital signage, and enterprise communications, I bring a perspective grounded in both philosophical rigor and practical execution.
I co-founded Automotive Broadcasting Network, a national SaaS media platform powering 12,000+ automotive endpoints across interactive systems, digital signage, and OEM networks, and led it through acquisition by Spectrio. As VP of Product Management at Spectrio, I spearheaded platform consolidation across multiple legacy systems, modernized our flagship CMS with AI-enhanced workflows and advanced analytics, and directed product strategy across the company's SaaS and digital-experience platforms.
Today, I’m building. Three products shipped in the last six months—a data intelligence platform, a job search tool in private beta, and a family storytelling game with a waitlist. I conceive, architect, and launch full-stack products from zero to production, and I help other teams do the same.
From napkin sketch to launch plan. Market positioning, feature prioritization, go-to-market strategy, and the hard calls about what to build and what to kill.
Full-stack product development from architecture through production deployment. Three products shipped in six months with no traditional engineering team.
Interim and fractional leadership for teams navigating platform consolidation, post-acquisition integration, or the messy middle between prototype and scale.
Multi-drive data intelligence platform that crosses every boundary between your scattered files and the answers trapped inside them. Scans local drives, cloud storage, email archives, and compressed bundles across Mac, Windows, and Linux. Extracts text from 30+ formats including documents, images (EXIF + OCR), and audio/video via local Whisper transcription. Every file gets relevance-scored 0–100 by a configurable keyword engine, deduplicated across sources, and made searchable through semantic embeddings. Ships with an MCP server (18 tools, 5 resources, 4 prompts) and REST API for builder integration. Local-first by default—cloud when you choose.
Python · SQLite · FastAPI · MCP Protocol · Whisper · OCR · Vector Embeddings · YAML Config
In development · troverknows.com →
Cooperative family storytelling game where an AI Game Master guides kids and parents through adventures that evolve based on how they play. iPad app with physical companion cards and storybooks. The AI remembers every choice, every character name, every brave decision—and weaves them into future stories. No screens-only isolation. No competitive scoring. Just families building a shared narrative that grows richer the more they play together.
AI Game Master · iPad · Physical + Digital Hybrid · Persistent Narrative Memory
Waitlist open · thevillageremembers.com →
Production AI-powered job search platform in private beta. Built entirely with AI agents as the engineering team—100+ features shipped across 20+ releases with zero traditional developers. The platform scores job matches against your resume and dream job profile using Claude, delivers personalized morning briefing emails, and searches 9 job board APIs simultaneously. Designed for senior product managers who are tired of keyword-matching job boards that waste their time.
Next.js · Supabase · Claude API · RAG pipeline · Multi-agent development · Vercel
Live in private beta · Join the Beta →
Governance framework for AI-assisted software development. Thirteen articles covering chain of command, mandatory planning, code quality enforcement, debugging hierarchy, prohibited actions, and amendment processes—every rule written in response to a real production failure. When you run AI coding agents against a production codebase, you need more than prompts. You need a constitution. Works with Claude Code, Cursor, Copilot, Windsurf, or anything that reads a CLAUDE.md.
Constitutional AI governance · Multi-agent coordination · Production-tested
Open source · View on GitHub → · axispraxis.ai →
MCP server for multi-agent coordination. When you run multiple AI coding agents against the same codebase, someone has to coordinate them—and it shouldn’t be you. The bridge provides structured checkpoints, per-agent guidance channels, auto-read hooks, and constitutional compliance checking. File-based, no database, no WebSocket—just JSON on disk that any agent can read. Born from running three Claude Code agents in parallel against a production app and nearly losing my mind.
MCP protocol · Multi-agent orchestration · Async coordination · File-based IPC
Open source · View on GitHub →
Transforms 20 years of Apple Mail archives into searchable intelligence using RAG architecture—463,000 emails, 11GB, 16 years of context made semantically searchable. Parses .emlx, reconstructs threads, extracts attachments, and builds a contact graph. The research that led directly to Trover’s email extraction pipeline.
RAG · ChromaDB · Vector embeddings · Apple Mail · Flask + Next.js
Production prototype · Request Access →
Agentic assistant for complex project management. Automates meeting intelligence, status aggregation, and cross-team dependency tracking while preserving the human role in risk assessment and stakeholder navigation. Built as a demo for enterprise program management teams managing 50+ concurrent workstreams.
Multi-agent orchestration · Claude · Function calling · Enterprise PM
Exploring pilot opportunities · Request Access →
Automated marketing platform for multi-location dealerships. Manages web properties, ad campaigns, and content generation across hundreds of rooftops while preserving local identity. Built on 15+ years of automotive industry experience from co-founding and scaling a national SaaS media platform to acquisition.
Image generation · Content synthesis · SEO · Multi-location management
Concept / discovery · Request Access →
Ingests virtual data rooms to perform structured due diligence—financial cross-referencing, contract clause extraction, risk flagging, and management team analysis. Designed for buy-side PE/VC teams who need to process hundreds of documents in days, not weeks. Grounded in firsthand experience building and selling a company.
Document parsing · Cross-reference mapping · Structured extraction · VDR ingestion
Concept / discovery · Request Access →
Not because it’s a game—because it’s a slot machine wearing a game costume. 20 years of building SaaS products taught me what a dark pattern looks like. So I’m building something different.
Read on LinkedIn →Most job platforms grow when your search drags on. Their incentive is your inefficiency. We built Last/JobScout with the opposite goal. If you leave, it means you landed. That’s the only metric that matters.
Read on LinkedIn →100+ features. 20+ releases. Zero traditional developers. The entire product was built using Claude Code agents governed by the Axis Praxis constitution. Every rule exists because something broke in production.
Read on LinkedIn →They exist because companies can extract optionality without obligation. A role can stay posted indefinitely; a candidate cannot. You weren’t unlucky—you were interacting with a system optimized for optionality, not resolution.
Read on LinkedIn →Vibe coding gets you a demo. Governance, memory, and conviction get you a product. What two months of production AI development taught me about the difference.
Read on LinkedIn →Gartner says 50% of companies that cut staff for AI will rehire by 2027. AI should also serve the people dealing with the consequences—not just the people making the cuts.
Read on LinkedIn →Introducing the Axis Praxis Constitution—a governance framework for teams building software with AI agents. 13 articles. Every rule exists because something broke.
Read on LinkedIn →An AI-first job search agent designed to restore clarity and trust to the job search process. Scores jobs against your actual resume. Flags ghost postings. Free private beta.
Read on LinkedIn →Reflecting on Andrew Ng's metaphor and Hume's is/ought problem—the power of AI lies not just in what it can do, but in how it sharpens the choices we face.
Read on LinkedIn →Exploring diffusion models through cosmic metaphor—what happens when AI learns to create the same way the universe did, sculpting order from chaos?
Read on LinkedIn →