Senior Data Scientist

**Judgement:** This skill is a strong, production-oriented toolkit for senior-level data science and experimentation. It encapsulates common workflows (A/B testing, feature engineering, ML evaluation, causal DiD) with sound statistical methods and sensible defaults. **Key strengths** - Solid statistical foundations: proper sample size calculation, two-proportion z-tests, DiD with HC3 robust errors. - Good engineering practices: train/test separation in pipelines, cross-validation with stratification, overfitting diagnostics, MLflow logging. - Embedded best-practice checklists that reflect real-world senior DS experience. **Risks / limitations** - Heavy reliance on the Python scientific stack (NumPy, SciPy, pandas, scikit-learn, XGBoost, statsmodels, MLflow); environments lacking these will need setup and may face dependency/version issues. - Limited defensive handling of edge cases (e.g., extremely low counts in experiments, degenerate class distributions) may require additional checks in critical production paths. - Causal inference coverage is focused on a basic DiD pattern; more complex designs (IV, RDD, advanced matching) are out of scope. **Recommended scenarios** - Mature product/analytics teams needing a standardized, reliable way to run and analyze A/B tests. - Data science groups building or refactoring tabular ML pipelines with best-practice evaluation and logging. - Organizations performing recurring policy or feature-impact assessments where quick, reproducible DiD analyses are sufficient. - As a reference implementation or starting point for building a broader internal experimentation and modeling framework.