Senior platform engineer, Linux kernel contributor, and author of the Cloud-V public RISC-V cloud platform and its RISC-V CI infrastructure. I work on Kubernetes, distributed heterogeneous computing, and bare-metal infrastructure at scale. I write about CI, platform engineering, the role of AI in them, and the tools I build.
CI logs are mostly noise. I built a tool that reads them for you.
A failed CI run dumps thousands of lines to tell you one thing went wrong. ci-medic distills the log, redacts secrets, and posts a verdict (what failed and why) where you already look. Here is how it does on 13 real failed logs.