i build backend systems and ai agents β the kind of software where the hard part is invisible and the bugs only appear after midnight. conveniently, that's when i start working.
off the clock i'm into paranormal investigation, which turns out to be the same job as debugging: something impossible happened, nobody witnessed it, and the logs are useless. i believe every haunting has a root cause. so far, it has always been a race condition.
also: cats. πββ¬
other people's codebases, other people's hauntings. 7 PRs merged across 6 upstream projects, three of them cncf.
- apicurio registry β #8684 merged. then found #8926: MCP tool annotations were being validated against the wrong schema entirely, so not one of the four hints anyone wrote was ever checked. fix in review β #9055.
- opentelemetry go compile-time instrumentation β #856 merged. then found #886: struct rules cheerfully matched things that were not structs, and inside a grouped
typeblock a field could land on a struct nobody named. fix in review β #887. - kubeflow pipelines β #13780 merged: the python sdk was quietly adding
--trusted-hostto every pip call. opt-in now. - microsoft/vscode β #324571 and #324734 merged in v1.129.0: a rerun-last-task regression, and a pinned tab that refused to let go when dragged.
- nodejs/node β #64365 Β· fedora-infra/bodhi β #6121
- in review: kserve #5908 Β· #5909 Β· textualize/textual #6631 Β· #6630
the half i like best: find it myself, report it, then fix it.
a competitive programming platform, built end to end β 95+ problems, an async code judge that speaks 10 languages, real-time verdicts over websockets, elo ratings, plagiarism detection.
the favourite child: a custom segment tree that can replay the leaderboard at any minute of any past contest. time travel, but auditable.
typescript Β· nestjs Β· next.js Β· postgresql Β· redis Β· bullmq β source
security incident response, run by two LLM agents keeping each other honest β a Commander proposes actions, an Oversight agent approves, revises, or vetoes.
both agents trained with GRPO reinforcement learning on qwen2.5-3b. placed top 100 at the meta pytorch Γ scaler openenv hackathon.
python Β· pytorch Β· grpo Β· docker β source
search-as-you-type, built like it has to survive production β a redis cache sharded across 3 nodes on a consistent-hash ring, an in-memory trie serving top-K suggestions, a batched writer that collapses thousands of db writes into a handful.
benchmarked properly: p50, p95, p99.
python Β· fastapi Β· redis Β· sqlite
a sales agent that haunts your google calendar, benevolently. thirty minutes before every meeting, it researches the prospect on the live web and emails you a brief with cited sources.
nine google cloud services stitched together, zero manual invocation. it just shows up, prepared β like you wish you did.
next.js Β· gemini Β· cloud run Β· firestore
a poltergeist you can hire. it drives a real chromium browser with actual mouse and keyboard events, finds form fields through the accessibility tree, and streams every action and screenshot to a live dashboard.
so you can watch the ghost work.
python Β· playwright Β· fastapi Β· docker β source
an internet speed test that doesn't try to sell you a VPN. download, upload, ping, jitter, bufferbloat β measured against cloudflare's edge with zero backend, rendered on a hand-rolled SVG speedometer.
react Β· vite Β· typescript Β· cloudflare pages β source
- global top 6 β apple developer academy, italy
- top 100 β meta pytorch Γ scaler openenv hackathon
typescript python go java c++ nestjs fastapi quarkus express react next.js websockets postgresql redis mongodb sqlite pytorch grpo gemini Β· claude Β· openai docker gcp cloudflare vercel linux
/\_/\ ( o.o ) > ^ <
if your codebase does something that can't be explained β i want to hear about it.

