blog
Notes from the build
Backend and data infrastructure mostly, plus whatever else I have been taking apart lately. Written when something took longer to work out than it should have — usually so I do not have to work it out twice.
- 28 Oct 20269 min
FFmpeg ate my server: lessons from a browser video pipeline
A beginner-friendly tour of containers, codecs, remuxing and transcoding — told through every way FFmpeg broke the interview pipeline I built, from a phantom 1,000 fps to five duplicate scores for one candidate.
- #ffmpeg
- #video
- #node
- #reliability
- 27 Oct 20267 min
Serving Markdown to AI agents: .md exports, llms.txt and the caching traps in between
How I gave a Django marketing site and a WordPress blog a clean, machine-readable twin of every page — and the CDN, ETag and redirect bugs that only show up once real agents start fetching.
- #ai
- #django
- #wordpress
- #http
- #caching
- 26 Oct 20266 min
100% CPU on WordPress: the sitemap nobody was caching
Two ECS tasks pinned at full CPU, mostly by scrapers asking for sitemaps. How I traced it to a two-year-old hotfix, why the SEO plugin's own cache couldn't be trusted, and the fingerprint-based cache that fixed it.
- #wordpress
- #aws
- #performance
- #caching
- 21 Apr 20254 min
Generating knowledge base articles from Jira tickets
Jira is a good tracker and a terrible knowledge base. A small Node service that turns resolved tickets into documentation drafts, and the parts of it that were harder than expected.
- #llm
- #node
- #mongodb
- #documentation