I don't just use AI — I build pipelines where AI ships real work.
Automated open-source contribution workflows, developer CLIs, and applied ML.
- AI-assisted open source — I run a daily AI-augmented pipeline that triages maintainer feedback, keeps pull requests healthy, and ships focused fixes upstream. Merged into projects like
pnpm,remotion,sentry-javascript,hydra,theia, and more — with active PRs acrossfastify,storybook,weblate,CyberChef, and friends. - Developer tooling — small, dependency-free CLIs that turn messy workflows into checks you can run in seconds.
- Applied ML & data — notebooks and deployed apps: price prediction, clustering, computer vision, and NLP experiments.
⚙️ Transparency: much of my daily OSS work is orchestrated by an AI routine I designed — the pipeline drafts, the human reviews, quality gates decide. Every contribution is tested and written to be easy to review.
| Project | What it is |
|---|---|
oss-pulse |
Live dashboard of my open-source activity — auto-refreshed daily, zero manual edits |
cache-compass |
Finds and safely removes stale dev caches across mixed-language repos |
doclink-auditor |
Audits Markdown links and anchors so broken docs never ship |
atomic-habits-chatbot |
RAG chatbot (AnythingLLM + Gemini) — live demo |
Rental-Price-Prediction-in-Istanbul |
Scraping → features → models → deployed Streamlit app |
Data-Science-Projects |
30+ ML/DL notebooks: tabular, CV, NLP, recommenders, time series |

