AI & Machine Learning
Start with the shared base, then follow the machine learning track or the AI track, and put it to work on real problems.
The shared base
Learn Python, probability, and the classical machine-learning ideas both tracks assume.
- 01
Python for Data Work
Turn Python records into trustworthy, analysis-ready tables with NumPy and pandas.
Open courseBefore thisNo prerequisite courses - 02
Probability and Statistical Reasoning
Build probability from cases, then use it to reason about repeated observations and evidence.
Open courseBefore thisPython for Data Work - 03
Classical Machine Learning
Take one prediction problem through data cleaning, modeling, deployment, and monitoring.
Open course
Machine learning
Model, evaluate, and decide: the ladder behind data scientist and ML engineer interviews. Follow each course’s prerequisites; the later ones are not strictly sequential.
- ↗
Feature Engineering
Turn raw data into useful ML features, with reliable splits, transformations, and model comparisons.
Open course - ↗
Collaborative Filtering Recommender Systems
Predict tastes and rank recommendations from a sparse ratings matrix using neighborhoods, latent factors, and honest evaluation.
Open courseBefore thisClassical Machine Learning - ↗
Experimentation and Product Analytics
Turn an advertising-product experiment into defensible metrics, comparisons, and a product recommendation.
Open courseBefore thisProbability and Statistical Reasoning - ↗
Causal Inference
Learn what causal inference means, why assumptions matter, and how to estimate treatment effects step by step.
Open course - ↗
Reinforcement Learning
Learn through trial and error: start with bandits, then teach a small robot to reach a goal with Q-learning.
Open coursePut it into practiceProject · Perishable Inventory Replenishment
AI engineering
Build reliable software around models: understand the model, retrieve evidence, use tools, evaluate behavior, and design the complete system.
- ↗
Software Engineering for AI Applications
Structure, test, and debug the software around a model instead of treating the model call as the whole product.
Open coursePut it into practicePair programming · Debug Report Chunking Before LlamaIndex - ↗
Deep Learning Foundations
Learn the computation underneath transformers before attention, fine-tuning, and serving add more machinery.
Open course - ↗
LLM Foundations
Understand what a language model actually computes, so the cost, latency, and failure modes of an LLM feature stop being surprises.
Open courseBefore thisDeep Learning FoundationsPut it into practicePair programming · Repair the LangChain RAG Tests - ↗
RAG and Retrieval Systems
Retrieve useful evidence, ground generated answers, and measure retrieval separately from generation.
Open course - ↗
Agents and Tool Use
Decide when a loop beats a pipeline, design tools a model can call correctly, and contain an agent that will eventually be wrong.
Open coursePut it into practicePair programming · Complete the CrewAI Tool Guard - ↗
Operating LLM Systems in Production
Evaluate LLM behavior across prompts, judges, RAG components, distribution shifts, guardrails, and long contexts.
Open coursePut it into practicePair programming · The Tuesday Prompt Tweak - ↗
Applied AI System Design
Design a complete model-backed product and defend its interfaces, measurements, trade-offs, and failure modes.
Open course - ↗
AI Security, Safety, and Privacy
Threat-model the complete AI system: its data, prompts, models, retrieval layer, tools, users, and logs.
Open coursePut it into practicePair programming · Complete the CrewAI Tool Guard
Model training and inference
Go deeper when the role expects training, alignment, serving internals, or multimodal systems.
- ↗
LLM Training, Data, and Scaling
Understand what pretraining consumes, how data quality changes the result, and why large-model training is a distributed-systems problem.
Open course - ↗
Post-training and Alignment
Learn how demonstrations, preference rankings, and carefully constrained updates turn a pretrained model into an assistant.
Open course - ↗
LLM Serving and Inference Economics
Find out where the latency and the money actually go, then change the number instead of guessing at it.
Open courseBefore thisLLM FoundationsPut it into practicePair programming · Repair the AI API Reliability Tests - ↗
Multimodal Models
Understand how models connect language with images, audio, and video—and where cross-modal systems fail.
Open course