Advanced Data Science
Building and Evaluating AI Systems
Open interactive chapters on how modern data and AI systems learn, fail, are evaluated, and connect evidence to responsible decisions.
Week 01Decisions
Choosing an AI System
When should we use a predictive model, a foundation model, or no AI?
Week 02Foundations
Neural Foundations
What actually changes during learning?
Week 03Foundations
Generalization
How do we know learning will transfer?
Week 04Foundations
Specialized Architectures
When does an architectural bias help?
Week 05Evaluation
Evaluation Before Application
What evidence would justify trust?
Week 06Representations
Embeddings and Attention
How do objects become useful vectors?
Week 07AI systems
Foundation Models
What evidence supports adaptation?
Week 08AI systems
Retrieval and Grounding
How can answers use inspectable sources?
Week 09Multimodal
Vision and Document AI
What is gained and lost across modalities?
Week 10Multimodal
Speech and Accessibility
How should speech transformations be evaluated?
Week 11Agents
Bounded Agents
When may a probabilistic model act?
Week 12Governance
Selection and Release
Which system should be released—and when?
Week 13Integration
Adversarial Testing
Can the system survive realistic failure?
Week 14Synthesis
Defense and Responsibility
What evidence supports the system?