yRocket

Wafer-level Zernike polynomials are used as spatial feature descriptors to decompose complex process variations, enabling polynomial-based machine learning models to accurately predict and optimize semiconductor manufacturing metrics such as yield and thickness.

A Python MLOps (Machine Learning Operations) sandbox and reference portfolio — building reproducible, production-ready machine-learning workflows and the operational engineering practices behind reliable model delivery.

Applied ML and statistics notebook for measurement-heavy engineering data — data profiling and modality taxonomy, machine-signal parameterization, feature engineering for wide time series, agreement metrics (CCC, CAI, R² vs MAPE), and piecewise regression.