Hands-on machine learning builds with full source code, detailed write-ups, and one-click Google Colab launchers. Every project walks through the theory, implements the model from scratch in Python, and benchmarks results on real datasets.
7 projects across 4 categories · Python · PyTorch · TensorFlow · scikit-learn · Jupyter notebooks · Google Colab
Build a GAN-powered super-resolution model that 4× upscales low-resolution images using SRGAN and PyTorch, then wire up a real-time face detector with OpenCV Haar Cascades. Both projects include full training pipelines, dataset links, and performance benchmarks.
Explore builds
Fine-tune a multilingual T5 model to summarize documents across multiple languages, then build a retrieval-based chatbot that answers questions using Wikipedia as its knowledge base via TF-IDF similarity. Each notebook explains the architecture choices and evaluation metrics.
View NLP projects
Apply both classical (ARIMA) and deep learning (LSTM) approaches to forecast S&P 500 index movements, and build a demographic model projecting Morocco's age-pyramid evolution through 2070 with Pandas and Matplotlib. Covers stationarity tests, feature engineering, and walk-forward validation.
View forecasts
Implement Q-learning from scratch to train a Tic-Tac-Toe agent through thousands of self-play episodes. Explore reward shaping, epsilon-greedy exploration, and how hyperparameter choices affect convergence speed. The notebook includes interactive visualizations of the learning process.
Open notebooksEach project is designed to be fully reproducible. You can read the write-up, study the code, and run the experiments yourself without installing anything locally.
Complete, commented Python notebooks hosted on GitHub. Clone the repo or browse inline.
One-click Google Colab launchers with pre-installed dependencies. Run experiments in your browser.
Direct links to every dataset used, with loading scripts and preprocessing steps included.
Theory and implementation walk side by side. Every design choice and hyperparameter is explained.