Projects & Ventures

A few live bets at the intersection of capital, operators, and AI.

Funds, platforms, and personal R&D. Some are mine, some are cofounded, one is a private project. The thesis under all of them: there are great businesses doing necessary work in unsexy industries, and the next generation of operators can run them better with modern tools.

Cofounder · LMM Private Equity

Search Fund Ventures

Lower-middle-market private equity firm investing in profitable, essential SMBs. We back the next generation of operators acquiring and running classic American businesses.

SFV invests in mission-critical small businesses doing necessary work in unsexy industries — and backs the operators acquiring them. The thesis is simple: there are great businesses already cash-flowing, and the next generation of operators can run them better with modern tools.

Insensitivity to market cycles. De-risked, well-known brands. Proprietary deal sourcing. Solid base financials. The possibility of AI-driven, tech-driven transformation of core products or operations. Operators with startup or entrepreneurial DNA.

We're looking for accredited investors with entrepreneurial experience to join our investor-advisor program — align your investing work with your career and help us help operators win.

searchfundventures.co
Cofounder · Investor Syndicate

SMB Investor Network

Syndication platform for accredited investors backing the operators acquiring profitable SMBs. $50K–$200K checks, hands-on support, real economics.

SMBIN is the syndicate side of the SFV stack — accredited investors writing $50K–$200K checks into SMB acquisitions, then going all-in: sharing expertise, making connections, providing hands-on support.

It isn't the big family-office checks that make these deals successful. It's smaller investors with skin in the game. The Stanford 2024 Search Fund Study shows roughly 30% IRR averages and 3–5× returns on the profitable deals — driven by operator support, not capital.

If you're a $200K+ earner curious about ETA, this is the on-ramp. Live deals, mentorship, and a community of operator-investors.

smbinvestornetwork.com
Head of R&D

ARCOS Labs

Enterprise IP protection for the AI era. Building the systems that let companies actually own their data and models in a world of pervasive AI.

ARCOS Labs is the parent entity for the AI-era IP protection work. We build the forensic and rights infrastructure that lets studios, publishers, and rights holders set the terms of engagement with generative AI companies.

VN.ai is the flagship product line.

arcoslabs.co
Forensic IP Intelligence

VN.ai

Studios, publishers, and rights holders use VN to set the terms of engagement with generative AI companies. Detect training. Grant consent. Control output. Audit everything.

Every major AI system trained on the open web, and your IP was in that data. VN proves it — then helps you decide what happens next.

Three layers of detection feed one unified evidence pipeline. Web detection flows into AI analysis, which flows into training-data review. The most complete forensic picture available, built to the rigor courts require.

From the forensic record, you can litigate, license, or both. The Forensic Suite produces a litigation-ready package; the Rights Suite turns the same infrastructure into licensing revenue with continuous enforcement and a tamper-evident audit trail.

  • 01 · Detect

    Prove which models trained on your work.

  • 02 · Consent

    Granular permissions for the AI era.

  • 03 · Control

    Set output policy on every license.

  • 04 · Audit

    A tamper-evident compliance trail.

vn.ai
Personal R&D

Metatransformer

Private project on voice-first human-computer interaction. Mic and camera in. Agents in the middle. Done in the world.

A private R&D project on voice-and-video first human-computer interaction. Mic and camera in. Agents in the middle. Clean handoffs to a small operator team for the last mile.

The keyboard is the bottleneck for a portfolio operator running a fund, multiple companies, and a family office at once. Most of what gets done in 2026 should not require sitting at a desk.

@metatransformr