Run A/B tests and causal inference

Build ML pipelines with feature engineering, hyperparameter tuning, cross-validation, and experiment tracking

Reproducible ML research where feature selection, experiment tracking, and cross-validation are non-negotiable.

• · dataset (CSV/Parquet)
• · model requirements (classification/regression/clustering)

• · trained model
• · evaluation metrics (F1/AUC/RMSE)
• · feature importance analysis

• · scikit-learn
• · pandas
• · PyTorch or TensorFlow
• · MLflow

Python 3.8+, labeled data, experiment tracking setup

• · Data leakage between train/test splits
• · Model overfits on small dataset

• · Cross-validation k-fold
• · Feature importance plots
• · MLflow experiment logs