- [Slides] Course Introduction; Foundations of Machine Learning
- [Assignment] Demand Forecasting at H&M
- [Web App] Demand Forecasting at H&M
- [Notebook] Linear Regression, Feature Engineering, and Regularization
- [Case] Demand Forecasting at H&M
- [Article] H&M, a Fashion Giant, Has a Problem (New York Times)
- [Article] H&M Pivots to Big Data (Wall Street Journal)
Science and Strategy of Artificial Intelligence
This course demystifies artificial intelligence for a non-technical audience. We go light on math and heavy on intuition. The goal is for you to build an appreciation for how underlying AI models work and how they unlock value from data in practice. We will explore various AI models using a mix of fully pre-written Python code and interactive web apps -- no prior coding experience is needed.
We will cover machine learning, reinforcement learning, neural networks, computer vision, and large language models (LLMs). We'll pay special attention to how LLMs are trained and aligned, and how the underlying technology shapes the economics and market structure of today's AI industry. We'll also build a tiny LLM along the way.
The course culminates in a product sprint where you'll build an AI-powered app deployed on the web, showcasing your skills and serving as a memento of the course.
Term Spring 2026 - 4 units - in person
Time Mondays, 1:00-3:50pm
Location G419
Instructor Auyon Siddiq
Product Sprint
Weekly Modules
- [Slides] Tree-Based Models: CART, Random Forests, XGBoost
- [Assignment] Range Rover Pricing
- [Web App] Range Rover Pricing
- [Notebook] XGBoost & Hyperparameter Tuning
- [Case] Algorithmic Pricing: Carvana and Zillow
- [Demo] Regression Tree Splits
- [Article] Zillow's Home Buying Debacle (CNN)
- [Slides] Reinforcement Learning
- [Assignment] Airline Revenue Management
- [Web App] Airline Revenue Management
- [Case] Adaptive Experimentation: Amazon and Alibaba
- [Demo] Bandits and Q-Learning
- [Article] Chart a Course for Reinforcement Learning (McKinsey)
- [Article] Starting Off on the Right Foot (Boston Dynamics)
- [Slides] Neural Networks
- [Assignment] Handwritten Digit Recognition
- [Web App] Handwritten Digit Recognition
- [Notebook] Neural Networks
- [Demo] Gradient Descent Simulation
- [Demo] TensorFlow Playground
- [Article] New Navy Device Learns by Doing (New York Times, 1958)
- [Article] Learning Representations by Back-Propagating Errors (Nature)
- [Slides] Computer Vision and Convolutional Neural Networks
- [Assignment] Image Recognition
- [Web App] Image Recognition
- [Notebook] Convolutional Filters
- [Notebook] Transfer Learning
- [Article] Inside Ralph Lauren's AI Styling Tool (Vogue)
- [Slides] Transformers and Large Language Models
- [Assignment] Yelp Review Generation
- [Web App] Yelp Review Generation
- [Notebook] Building a Tiny LLM
- [Demo] Embedding Projector
- [Slides] LLM Training and Economics
- [Assignment] 10-K Due Diligence
- [Web App] 10-K Due Diligence
- [Data] 10-K Excerpts
- [Article] How Scaling Laws Drive Smarter, More Powerful AI (NVIDIA)
- [Slides] Building LLM Apps
- [Kit] App Development Kit
- [Kit] App Development Kit
- [Office Hours] Product Sprint Check-In
Evaluation
Teaching Team