Machine learning

Steps

Foundations

1. 1 The theory behind learningReview the linear algebra, probability and optimisation behind supervised learning so later algorithms read as variations on familiar ideas.

   • Stanford CS229: Machine Learning cs229.stanford.edufree

     Course materials, lecture notes, and syllabus for Stanford University's flagship academic machine learning course.

   I finished this step
2. 2 Your first modelsFit, predict and score simple models with a consistent API, learning the train and test workflow every project follows.

   • Scikit-Learn User Guide and Tutorials scikit-learn.orgfree

     Comprehensive documentation and guided educational examples covering classical machine learning algorithms.

   • Machine Learning Mastery machinelearningmastery.comfree

     A popular technical tutorial blog offering practical, code-focused guides for applied machine learning.

   I finished this step
3. 3 Datasets for practicePick well-documented benchmark datasets to experiment with, so you can compare your results against known baselines.

   • UCI Machine Learning Repository archive.ics.uci.edufree

     A prominent academic archive providing benchmark empirical datasets for evaluating predictive and classification algorithms.

   • OpenML openml.orgfree

     An open, collaborative scientific platform and database for sharing machine learning datasets, tasks, and empirical experiments.

   • Kaggle Datasets kaggle.comfree

     Community platform offering thousands of open datasets, code notebooks, and benchmarks across business analytics and predictive modeling.

   I finished this step

Classical machine learning

1. 4 Linear models and regularisationUnderstand linear and logistic regression, overfitting and regularisation, the ideas that explain why models generalise or fail.

   • Scikit-Learn User Guide and Tutorials scikit-learn.orgfree

     Comprehensive documentation and guided educational examples covering classical machine learning algorithms.

   • Stanford CS229: Machine Learning cs229.stanford.edufree

     Course materials, lecture notes, and syllabus for Stanford University's flagship academic machine learning course.

   I finished this step
2. 5 Evaluation and validationUse cross-validation, suitable metrics and data leakage checks so reported accuracy reflects how a model behaves on new data.

   • scikit-learn scikit-learn.orgfree

     A Python library featuring efficient tools for predictive data analysis, classification, regression, and clustering.

   • Machine Learning Mastery machinelearningmastery.comfree

     A popular technical tutorial blog offering practical, code-focused guides for applied machine learning.

   I finished this step
3. 6 Trees and gradient boostingLearn decision trees and boosted ensembles, often the strongest choice for tabular data, and train them with dedicated fast libraries.

   • XGBoost xgboost.readthedocs.iofree

     An optimized distributed gradient boosting library designed for efficient and scalable machine learning.

   • LightGBM lightgbm.readthedocs.iofree

     A fast, distributed, high-performance gradient boosting framework based on decision tree algorithms.

   I finished this step
4. 7 Hyperparameter tuning (optional)Search model settings automatically instead of by hand, with pruning to stop unpromising trials early and save compute.

   • Optuna optuna.orgfree

     An open-source hyperparameter optimization framework specifically engineered for machine learning and deep learning models.

   I finished this step

Deep learning

1. 8 How neural networks learnBuild intuition for layers, gradient descent and backpropagation before writing code, so training behaviour makes sense when it goes wrong.

   • 3Blue1Brown Neural Networks 3blue1brown.comfree

     Animated video lessons explaining neural networks, gradient descent, backpropagation and transformers visually, suitable for learners with basic mathematics.

   • Introduction to Deep Learning sebastianraschka.comfree

     Free university-level video lecture series on deep learning, from neural network basics to convolutional networks, transformers and generative models.

   I finished this step
2. 9 Training with PyTorchWrite tensors, models and training loops in a deep learning framework, and learn to load data and track losses on a GPU.

   • PyTorch Tutorials docs.pytorch.orgfree

     Official collection of PyTorch tutorials and recipes covering tensors, training loops, vision, text and deployment for people learning deep learning in Python.

   • PyTorch pytorch.orgfree

     An open-source machine learning framework widely used in data science research for deep learning and neural network training.

   I finished this step
3. 10 Practical deep learningTrain image and text models end to end with transfer learning, focusing on what works in practice before diving into theory.

   • Practical Deep Learning for Coders course.fast.aifree

     A free, hands-on online course teaching deep learning and practical model deployment from top down.

   I finished this step
4. 11 Transformers (optional)Understand attention and the transformer architecture behind modern language models, then fine-tune pretrained models for your own tasks.

   • The Illustrated Transformer jalammar.github.iofree

     Heavily illustrated walkthrough of the transformer architecture, explaining attention, encoders and decoders visually for students new to modern language models.

   • Hugging Face NLP Course huggingface.cofree

     A hands-on, open-access course teaching practical transformer workflows, fine-tuning, and model deployment.

   I finished this step

Practice and research

1. 12 Tracking experimentsLog parameters, metrics and model versions for every run so experiments are reproducible and easy to compare later.

   • MLflow mlflow.orgfree

     An open-source platform to manage the machine learning lifecycle, including experimentation, reproducibility, and deployment.

   I finished this step
2. 13 Reading research (optional)Read clear visual explanations and open peer-reviewed papers to follow how the field evolves beyond tutorials and courses.

   • Distill distill.pubfree

     An online publication dedicated to clear and interactive visual explanations of machine learning and statistical models.

   • OpenReview openreview.netfree

     An open peer-review repository and scientific database managing publications and reviews for top AI conferences.

   • Journal of Machine Learning Research jmlr.orgfree

     An open-access peer-reviewed journal publishing research on algorithms, statistical theory, and empirical data science.

   I finished this step