Open Source
Open source contributions come in many forms: bug reports that survive for years, documentation gaps caught through real-world use, and code that gets merged into projects used by thousands of engineers.
The best place to start for a full listing of my activity is a search by author:
A fix I needed was already merged to cadvisor master, but stuck in no released tag for two months, quietly crash-looping kubelet on ZFS-graph-driver nodes. This is the story of moving it: the GitHub issues, the escalation that actually worked, the release that shipped ~90 minutes after a Slack ping, and the Kubernetes vendor-bump PR I opened that merged two hours later.
Edge0's founder is pulling 3,600+ likes on LinkedIn with a slick clip of an iPhone running a 35B model on-device. The trouble: there is no iOS app in the Edge0 repo, the founder concedes the phone app is 'coming soon,' and the demo reads as AI-generated. Meanwhile Silo, a one-developer app on the App Store since early 2025, already does private, offline, on-device LLM chat for real, with a year of unglamorous commit history to prove it. I walk the two repos side by side: what ships vs. what's narrated, and why the harder claim (35B MoE streaming on a phone) is still unproven.
antirez's h3.c runs a full MiniMax-H3 video model on Apple Silicon in ~35,000 lines of C, Objective-C, and Metal -- no Python, no PyTorch, no CUDA-style runtime. I compare it to the pain of getting PyTorch running on AMD CPU + AMD GPU + ROCm on Linux in mid-2023, and argue both stories are really about one thing: removing abstraction layers you don't actually need.
Discovered a gap in the Tailscale Kubernetes Operator through real-world GKE usage. Filed an issue and submitted a PR that was merged upstream. Tailscale has raised $160M in Series C funding -- yet this was missed by their team.
Work in progress: adding dedicated file-based logging for tailscaled on Debian/Ubuntu. Validated across 4 systems. Upstream PR to Tailscale (a $160M Series C company) not yet submitted -- but the gap is real, documented, and reproducible.
Ongoing work: developing safety scripts and APT hooks to protect ZFS pools during kernel upgrades on Armbian ARM64 systems. Found and fixed two critical bugs that left systems in an unsafe-to-reboot state after routine apt upgrade operations.
In February 2020 I reported a private image pull failure in the then-new GitLab Container Registry. Four years later the issue is still open, and engineers worldwide have used it as their reference.
Early adopter and community contributor across Ubuntu Linux, OpenSolaris, and OpenIndiana -- including production ZFS deployments on OpenIndiana before ZFS on Linux reached maturity.