Skip to content
ML to LLMs

From linear regression to LLM agents in production.

A visual, math-first course on how ML and AI systems work. Go deep, skim in quick-read mode, or prepare for a specific interview loop.

Start learning Open the labs

Skill progress

  • Mathematics0/6 lessons · 0%
  • Classical ML0/21 lessons · 0%
  • NLP0/4 lessons · 0%
  • Deep learning0/31 lessons · 0%
  • Transformers0/4 lessons · 0%
  • LLMs0/19 lessons · 0%
  • AI engineering0/11 lessons · 0%

Five ways to use it

Deep dive96 lessons in 30 modules. Intuition, an interactive lab, the math with every step, code, and engineering trade-offs.Full course Quick readFlip any lesson to Quick read: intuition, the lab, the summary, and interview questions in about a quarter of the time.Pick a lesson Watch it movePress Watch on any lab for a captioned, animated tour, or scroll an explainer where the lab moves as you read.Visual explainers Interview prepSeven-day sprints for ML engineer, AI engineer, applied scientist, and data scientist loops, plus a rapid-fire drill.Choose your role Quick referenceCheat sheets, model cards, the glossary, and the Math Lab when you need one fact fast.Cheat sheets

Interview sprints: ML Engineer · AI Engineer · Applied Scientist · Data Scientist · Rapid-fire drill

How everything connects

Every model in this course is a step in the same loop. Data becomes a representation, a model makes a prediction, a loss scores it, optimization improves it, and evaluation decides whether it ships. Click any stage or branch to open its lesson.

  1. 1Datacollect, clean, split
  2. 2Representationfeatures, vectors, tokens
    NLPDimensionality reduction
  3. 3Modela parameterized function
    RegressionClassificationClusteringNeural networksTransformers
  4. 4Predictionŷ for one input
    LLM applications
  5. 5Losshow wrong, as one number
  6. 6Optimizationfollow the gradient
    Backpropagation
  7. 7Evaluationmetrics on unseen data
    LLM evaluation
  8. 8Deploymentserve predictions
    Retrieval-augmented generationAgents and tool use
  9. 9Monitoringdrift, quality, cost
    Safety and security

Interactive labs

32 labs across 7 areas. Every number on screen is computed live, in your browser.

FoundationsUnderfitting vs overfittingChange polynomial degree, training size, and noise; watch training and validation error split apart.Neural Network LabGradient descentStep across a loss surface with SGD, momentum, or Adam, and push the learning rate until it diverges.Transformer LabSelf-attentionStep through Q, K, V, QKᵀ, scaling, the causal mask, softmax, and the weighted sum, with shapes.LLM and RAG LabRAG pipelineChunk documents, retrieve and rerank for a question, and see what happens when retrieval misses.LLM and RAG LabKV cache and inference memorySize the weights and KV cache for a model, context, and batch; see whether it fits and how fast decoding can go.AI EngineeringAgent loopStep through an agent's think, call tool, observe loop with mocked tools; inject failures and step limits.

Recently viewed

Prefer plain markdown?

The full course, with the same lessons, figures, runnable Python, quizzes, and model cards, reads directly on GitHub. No build step, no JavaScript.

Open the GitHub edition

The full path

30 modules, 96 lessons.

  1. 0Mathematical and programming foundations6 lessons
  2. 1ML vocabulary and learning paradigms2 lessons
  3. 2The end-to-end ML workflow2 lessons
  4. 3Regression2 lessons
  5. 4Classification2 lessons
  6. 5Instance-based and probabilistic models3 lessons
  7. 6Trees and ensembles2 lessons
  8. 7Unsupervised learning2 lessons
  9. 8Dimensionality reduction1 lesson
  10. 9Classical NLP4 lessons
  11. 10Neural-network foundations3 lessons
  12. 11Gradient descent and backpropagation2 lessons
  13. 12Training and regularization2 lessons
  14. 13Deep-learning architectures6 lessons
  15. 14Transformers4 lessons
  16. 15Large language models7 lessons
  17. 16Retrieval-augmented generation2 lessons
  18. 17LLM evaluation1 lesson
  19. 18LLM inference serving2 lessons
  20. 19Agents and tool use5 lessons
  21. 20Production AI engineering2 lessons
  22. 21Safety and security3 lessons
  23. 22ML system design and experimentation3 lessons
  24. 23Projects and career roadmap2 lessons
  25. 24Applied ML in practice5 lessons
  26. 25Computer vision and multimodal6 lessons
  27. 26Reinforcement learning5 lessons
  28. 27Training at scale5 lessons
  29. 28Prompting and context engineering3 lessons
  30. 29Speech and audio2 lessons