The thesis
Alpicat isn't a fund or a single business. It's the umbrella entity that holds stakes in and runs operational support for everything else, and more importantly it's where the shared machinery lives. The bet is simple: if one operator builds the right system once, that system can carry a whole portfolio. Automation built for one venture lowers the headcount needed to scale all of them. Specific knowledge plus leverage beats a bigger team.
What I do here
At the Holdings level I run cross-company strategy and build the tooling every venture reuses. In each individual company I wear a different functional hat — operations lead in one, investment structurer in another, revenue-ops engineer in a third, AI-automation builder in a fourth. Alpicat is the layer that makes wearing all of those hats at once actually possible without dropping threads.
What's built
The real product is the operating infrastructure:
- ◦Company "brains" — every venture gets a standardized, version-controlled knowledge base holding its context, operations, decision log, and active projects, synced on a schedule. Nothing important lives in my head.
- ◦An agentic workflow — I run my work through AI agents (Claude Code) wired into each company's real tools via MCP integrations, with a company-registry routing table so work always lands in the right company's context with only that company's credentials and tools.
- ◦Mission Control — a cross-company command center that launches an AI working session for any company, holds multiple live terminals per company, and aggregates open items from each venture's real source of truth onto one screen.
- ◦A communications monitor — a two-tier system that triages inbound across companies every thirty minutes and drafts replies behind a human approval gate.
- ◦A self-hosted homelab — scheduling, password/secret management, uptime monitoring, and a private mesh network tying my machines together, so the whole environment stays private, cheap, and under my own control.
- ◦A shared agent-template and skills library — reusable capabilities I can drop into a new company on day one.
Where it stands
Active and in daily use across every other star in the constellation. The shared-infrastructure layer is mature; the work now is compounding it — every new automation built for one company quietly upgrades the rest.
Nearby stars
Everything orbits here. Orchidea and El Trejo inherit the same brains and tooling on the energy side; MentorHood and Pass Gallery on the people-and-platforms side; Blueprint and SYNRGY on the sport side; GuzPicad on the frontier. Same operator, same operating system, radically different domains.