Build real ML models from linear regression to neural networks. Work on Kaggle datasets, tune models, and deploy predictions as APIs.
Implement gradient descent manually. Compare to scikit-learn. Plot cost curve.
Multiple regression on Boston/Ames dataset. Feature selection + evaluation.
Classification on Titanic. Feature engineering + Random Forest. Kaggle submit.
Binary classification. Handle class imbalance with SMOTE. ROC-AUC > 0.85.
Logistic regression on lending data. Threshold tuning for precision vs recall.
+15 more projects available after enrollment
Build a real project in 4 weeks