The AI knowledge tree
Every branch of AI except computer vision, as structured courses only: no articles, no blog posts, no "read this repo". Ten tiers ordered by dependency, each one stating what it builds on. Within a tier the domains are siblings and can be taken in any order.
Every entry names the course, the institution and its course code, so it resolves to exactly one thing. Nothing here is pruned to a shortest path: the tree covers each branch as it exists. Computer vision is deliberately out of scope, though the multimodal material in tier 5 necessarily touches it.
Tier 0. Mathematical foundations
Everything downstream assumes this material. Probability and statistics carry the most weight later.
Linear algebra
-
Linear Algebra (18.06)
-
Mathematics for Machine Learning: Linear Algebra
-
Matrix Methods in Data Analysis, Signal Processing & ML (18.065)
Calculus & matrix calculus
-
Multivariable Calculus (18.02)
-
Mathematics for Machine Learning: Multivariate Calculus
-
Matrix Calculus for Machine Learning and Beyond (18.S096)
Probability & statistics
-
Pick one
-
Statistics 110: Probability
-
Probability - The Science of Uncertainty and Data (6.431x)
-
-
MicroMasters in Statistics and Data Science
-
Bayesian Statistics: From Concept to Data Analysis
Information theory
-
Information Theory, Pattern Recognition and Neural Networks
-
Information Theory (EE276)
Optimization
-
Convex Optimization I (EE364A)
-
Convex Optimization II (EE364B)
-
Introduction to Numerical Methods (18.335)
Tier 1. Classical machine learning
The pre-deep-learning core: estimators, generalisation, and the theory the rest leans on.
Applied / intro ML
-
Machine Learning Specialization (3 courses)
-
Learning From Data
-
Pick one
-
Machine Learning (CS229)
-
Introduction to Machine Learning (10-601 / 10-701)
-
Statistical learning
-
Statistical Learning with Python
-
Statistical Learning Theory and Applications (9.520)
-
Machine Learning Theory (CS229M / STATS214)
Bayesian & probabilistic methods
-
Statistical Rethinking
-
Bayesian Methods for Machine Learning
-
Probabilistic Machine Learning
-
Numerics of Machine Learning
Probabilistic graphical models
-
Pick one
-
Probabilistic Graphical Models (CS228)
-
Probabilistic Graphical Models Specialization (3 courses)
-
-
Probabilistic Graphical Models (10-708)
Causal inference
-
Introduction to Causal Inference
-
Causal Inference & Causal Inference 2
Tier 2. Deep learning core
Backprop, optimisation, architectures, generative models.
Deep learning fundamentals
-
Introduction to Deep Learning (6.S191)
-
Pick one
-
Deep Learning Specialization (5 courses)
-
Practical Deep Learning for Coders
-
-
Deep Learning (CS230)
-
Pick one
-
Introduction to Deep Learning (11-785)
-
Designing, Visualizing and Understanding Deep Neural Networks (CS182)
-
-
Deep Learning
Generative models
-
Deep Generative Models (CS236)
-
Deep Unsupervised Learning (CS294-158)
-
Diffusion Course
Graph & geometric learning
-
Machine Learning with Graphs (CS224W)
-
Geometric Deep Learning
Tier 3. NLP, transformers & LLMs
From word vectors to building a language model end to end. CS336 is the densest item in the tree.
NLP foundations
-
Pick one
-
NLP with Deep Learning (CS224N)
-
Advanced NLP (11-711)
-
-
Natural Language Processing Specialization (4 courses)
-
LLM Course (formerly NLP Course)
Transformers & building LLMs
-
Neural Networks: Zero to Hero
-
Language Modeling from Scratch (CS336)
-
Transformers United (CS25)
-
Pick one
-
Generative AI with Large Language Models
-
Generative AI Engineering with LLMs Specialization
-
Fine-tuning & post-training
-
Finetuning Large Language Models
-
Reinforcement Learning from Human Feedback
-
Post-training & alignment (within CS336)
Prompting, RAG & context
-
ChatGPT Prompt Engineering for Developers
-
Prompt Engineering Specialization
-
Building and Evaluating Advanced RAG
-
Text Retrieval and Search Engines
Agents & tool use
-
AI Agents Course
-
MCP Course
-
Large Language Model Agents (MOOC)
Evaluation & interpretability
-
Evaluating and Debugging Generative AI
-
Automated Testing for LLMOps
Tier 4. Reinforcement learning
A standalone branch. Feeds post-training (RLHF, GRPO) back into tier 3.
Classical RL
-
Pick one
-
Reinforcement Learning Specialization (4 courses)
-
Reinforcement Learning Course
-
Deep RL
-
Deep RL Course
-
Spinning Up in Deep RL
-
Reinforcement Learning (CS234)
-
Deep Reinforcement Learning (CS285)
Tier 5. Other modalities & applications
Speech, tabular data, time series, and multimodal systems. Multimodal work necessarily touches vision.
Speech & audio
-
Audio Course
-
Spoken Language Processing (CS224S)
Time series & forecasting
-
Practical Time Series Analysis
-
Sequences, Time Series and Prediction
Recommenders & information retrieval
-
Recommender Systems Specialization (5 courses)
-
Mining Massive Datasets (CS246)
Embodied & robotics
-
Robotics Course (LeRobot)
-
ML for Games Course
Tier 6. ML systems & efficiency
Kernels, quantisation, distributed training, and serving. Where systems engineering meets models.
Parallel & GPU programming
-
Pick one
-
Parallel Computer Architecture and Programming (15-418/618)
-
Parallel Computing (CS149)
-
-
Fundamentals of Accelerated Computing with CUDA C/C++
Deep learning systems & compilers
-
Deep Learning Systems: Algorithms and Implementation (10-414/714)
-
Machine Learning Compilation
Efficient inference & TinyML
-
TinyML and Efficient Deep Learning Computing (6.5940)
-
Hardware Architecture for Deep Learning (6.5930)
ML systems design
-
Machine Learning Systems (CS249r)
-
Stanford MLSys Seminar Series
Distributed systems substrate
-
Distributed Systems (6.824/6.5840)
-
Intro to Database Systems (15-445/645)
Tier 7. MLOps & production engineering
Shipping, serving, and monitoring models once they leave the notebook.
MLOps
-
Machine Learning Engineering for Production (MLOps) Specialization (4 courses)
-
MLOps Zoomcamp
-
Full Stack Deep Learning
-
Made With ML
LLMOps
-
LLMOps
-
Full Stack LLM Bootcamp
Data engineering
-
IBM Data Engineering Professional Certificate
-
Data Engineering Zoomcamp
Cloud ML certifications
-
AWS Certified Machine Learning Engineer - Associate (MLA-C01)
-
Preparing for Google Cloud Certification: Machine Learning Engineer
Tier 8. Safety, security, privacy & governance
Alignment, adversarial robustness, privacy, and policy. Thin on formal courses, which is itself informative.
AI safety & alignment
-
AI Alignment
-
Frontier AI Governance
-
ARENA - Alignment Research Engineer Accelerator
Adversarial ML & LLM security
-
Red Teaming LLM Applications
Privacy-preserving ML
-
Privacy-Preserving AI series
Fairness, ethics & policy
-
AI Ethics / Responsible AI
-
EU AI Act compliance training
Tier 9. Research practice
Craft rather than knowledge: reading a field, reproducing results, and writing them up.
Doing and writing research
-
Writing in the Sciences
-
How to Write and Publish a Scientific Paper