Test & CI assurance for AI-generated codebases. Your build is green. I measure whether that green means anything.
Evidence rather than claims:
A release step in a Rust project could go green on a 404 and record the SHA of the error response. I reported it; the maintainer wrote and merged a fix within three days: <a href="https://github.com/unhappychoice/gitlogue/issues/236" rel="nofollow">https://github.com/unhappychoice/gitlogue/issues/236
My own CI once ran 58 test scenarios, crashed before writing any result, and the step still reported success, because continue-on-error hides more than you expect. Write-up: <a href="https://dev.to/onurkesim/do-not-trust-the-green-checkmark-48i5" rel="nofollow">https://dev.to/onurkesim/do-not-trust-the-green-checkmark-48...
My own tool's sabotage script and mutant files are public: <a href="https://github.com/onur-kesim/hafiza-kur" rel="nofollow">https://github.com/onur-kesim/hafiza-kur (faz0/)
A docs patch of mine is in drift's develop branch, committed by the maintainer: <a href="https://github.com/simolus3/drift/commit/d9f88e5b27531cfcfcd198e0e440963198175801" rel="nofollow">https://github.com/simolus3/drift/commit/d9f88e5b27531cfcfcd...
Caveats: 23 years in technical acceptance and maintenance of high-tech systems, including auditing a quality management system, but months, not years, in software. My code is mostly written by AI (Claude Code); the design, the checks and the verdicts are mine. I have not worked inside a team codebase or run a live system with real users. What I offer is a measurement, not seniority.
Looking for: part-time remote work (10-20 hrs/week) or fixed-scope audits of test suites and CI gates.
onurkesim · · focus · HN ↗
Remote: Yes (remote only)
Willing to relocate: No
Technologies: GitHub Actions (CI), Python, shell (CI steps), mutation testing (mutants checked into the repo), Claude Code (agentic workflows)
Résumé/CV: on request | GitHub: <a href="https://github.com/onur-kesim" rel="nofollow">https://github.com/onur-kesim | Writing: <a href="https://dev.to/onurkesim" rel="nofollow">https://dev.to/onurkesim
Email: onurkesimbjk@gmail.com
Test & CI assurance for AI-generated codebases. Your build is green. I measure whether that green means anything.
Evidence rather than claims:
A release step in a Rust project could go green on a 404 and record the SHA of the error response. I reported it; the maintainer wrote and merged a fix within three days: <a href="https://github.com/unhappychoice/gitlogue/issues/236" rel="nofollow">https://github.com/unhappychoice/gitlogue/issues/236 My own CI once ran 58 test scenarios, crashed before writing any result, and the step still reported success, because continue-on-error hides more than you expect. Write-up: <a href="https://dev.to/onurkesim/do-not-trust-the-green-checkmark-48i5" rel="nofollow">https://dev.to/onurkesim/do-not-trust-the-green-checkmark-48... My own tool's sabotage script and mutant files are public: <a href="https://github.com/onur-kesim/hafiza-kur" rel="nofollow">https://github.com/onur-kesim/hafiza-kur (faz0/) A docs patch of mine is in drift's develop branch, committed by the maintainer: <a href="https://github.com/simolus3/drift/commit/d9f88e5b27531cfcfcd198e0e440963198175801" rel="nofollow">https://github.com/simolus3/drift/commit/d9f88e5b27531cfcfcd...
Caveats: 23 years in technical acceptance and maintenance of high-tech systems, including auditing a quality management system, but months, not years, in software. My code is mostly written by AI (Claude Code); the design, the checks and the verdicts are mine. I have not worked inside a team codebase or run a live system with real users. What I offer is a measurement, not seniority.
Looking for: part-time remote work (10-20 hrs/week) or fixed-scope audits of test suites and CI gates.