MachineLearning
4 items
Series Posts
Part 0A: Neural Networks & The Learning Mechanism
Building deep intuition from first principles — how neural networks actually learn, explained step by step for engineers who want to truly understand the machinery before architecting with it.
Part 0: Python for Java Developers: From Streams to Tensors
A 10-part, example-driven Python series for seasoned Java and C# developers moving into data science, ML and AI. Follow Arjun, a Java architect pulled into a fraud-detection ML project, from culture shock to PyTorch.
Part 7: "Thinking in Arrays" — NumPy, Vectorisation and Broadcasting
Why the for loop is the enemy of numeric Python. The ndarray and its memory model, dtypes, vectorised operations, axes, boolean masks, broadcasting rules, views vs copies, NaN handling, random generators, and cosine-similarity search from scratch.
Part 9: "The First Model" — scikit-learn, Pipelines and Honest Metrics
Sentinel's first fraud model. The estimator API, stratified train/test splits, Pipeline and ColumnTransformer, logistic regression vs random forests, why accuracy lies on imbalanced data, precision-recall and threshold tuning, cross-validation, data leakage, and persisting models.