Local-first AI agents
Running Llama/Gemma models on-device with Apple FoundationModels / Ollama for privacy-preserving agent workflows
INITIALIZING
0%
Current engineering focus, active investigations, and what I'm building — updated weekly.
Technical deep-dives currently in progress — not yet production-ready.
Running Llama/Gemma models on-device with Apple FoundationModels / Ollama for privacy-preserving agent workflows
CRDTs for collaborative editing in agent workspaces — Yjs vs Automerge vs custom
FoundationDB-style simulation for distributed agent coordination logic
Spin/Wasmtime for portable agent plugins with capability-based security
Last 5 entries from the build log.
The homepage telemetry panel now reports real latency, uptime and GitHub activity instead of simulated bars.
Added a dedicated contact strip with mail, GitHub and socials — reachable without hunting through the footer.
Pagefind-backed search across posts, notes and research — wired into the nav console.
The terminal boots on /terminal — a full fake-OS with commands, games and man pages.
Select engagements where I can contribute architecture-to-operations ownership.
END OF NOW PAGE