Open-source contributor working across time-series forecasting, machine learning, and systems / command-line tools. I like small, correct changes backed by tests: GNU parity for Rust CLI utilities, and numerical correctness for Python ML and forecasting libraries.
- 🔭 Contributing to: uutils (coreutils, findutils), and Python ML/forecasting (darts, sktime, tslearn, torchmetrics, xgboost).
- 🧰 Languages: Python, Rust, TypeScript/JavaScript.
- 🧠 Interests: time-series forecasting, computational statistics, Bayesian modeling, distributed systems, LLM/AI tooling, parsers and CLI correctness.
- 🎓 CS background (ITAM): machine learning, algorithms, distributed systems, compiler design.
- uutils/coreutils —
uniqrepeated-D(#13898),foldGNU error messages (#14056),join -ofield accumulation (#14070) - uutils/findutils — non-UTF-8 argv (#833),
-sizeprefix (#834), signed-maxdepth(#835), ISO 8601 dates in-newerXt(#847), doubled error prefix on-type(#849) - dmlc/xgboost —
plot_importancevalue-label offset for small importances (#12497) - sktime/sktime —
mean_squared_log_errorcleanup (#10816) - modelscope/ms-swift —
format_timerounding (#9919) - unit8co/darts —
extract_subseriesreturn contract (#3184) - tslearn-team/tslearn — LCSS Sakoe-Chiba constraint (#705)
- 🤝 Open to interesting open-source collaborations in ML, forecasting, and systems tooling.

