AI-Powered Search authors Trey Grainger and Doug Turnbull are excited to offer an immersive virtual course on AI Search running November 4 to December 5th, 2025. Learn live with top experts and peers from top companies like Uber, Airbnb, AWS, Yahoo, Yelp, Wayfair, Shipt, Doordash, and more!

The Course

The AI-Powered Search: Modern Retrieval for Humans and Agents course is hands-on (with labs on most major topics), features many industry leading guest teachers (see below), and delivers multiple live class sessions with Doug and Trey every week.

All sessions (about 20 hours in total) will include live learning along with students from many top companies like Uber, AWS, Airbnb, Yahoo, Yelp, Wayfair, Shipt, Doordash, and more. In addition to the course content, Doug and Trey will also hold weekly office hours available exclusively to students in the course.

While you’re encouraged to attend live to maximize your learning opportunity, all students will have permanent access to all course recordings, the code / notebooks, slides, and a dedicated community channel with other students and with the instructors.

Expert Instructors at the Frontier of AI Search

The course is PACKED with leading experts in the field, helping you understand and put into practice the latest and greatest in AI Search techniques.

Course Leads

Expert Guest Instructors

Lightning Lessons

In the lead up to the course, Trey and Doug have been holding a series of FREE Lightning Lessons with top search experts. (By attending a Lightning Lesson, you’ll also get access to a promo code for a course enrollment discount):

If you like this kind of content, please support Doug and Trey by enrolling in the full course, where you’ll significantly improve your AI Search skills to build better, more relevant search for people and agents.

Course Syllabus

The full syllabus can be found on the course sign up page, but the high-level course topics are summarized below:

Week 1

Nov 4—Nov 7

• Course Overview + The Search Relevance Problem
  • 📄 Welcome & Course Overview
  • 📄 Language Models, Knowledge Graphs, and Knowledge Representation
  • 📄 Matching & Ranking
  • 📄 Lexical search
  • 📄 Embeddings & vector search
  • 📄 Dimensions of user intent
• User Behavior Insights (with Eric Pugh – OpenSource Connections)
• Signals & Reflected Intelligence Models
  • 📄 Collecting Proper Clickstream Data
  • 📄 Signals Boosting Models
  • 📄 Personalized Search & Behavioral Embeddings
  • 📄 Knowledge Graph Learning
  • 📄 Machine Learned Ranking (LTR) Intro

Week 2

Nov 11—Nov 14

• AI-Powered Query Modalities
  • 📄 Bi-encoders vs. Cross encoders
  • 📄 Multimodal Search
  • 📄 Hybrid Search
  • 📄 Intro to RAG
  • 📄 Late Interaction (Colbert, ColPali, etc.)
  • 📄 Combining Query Modalities (MiniCOIL, etc.)
• Mixing Sparse & Dense Representations with MiniCOIL (with Evgeniya Sukhodolskaya – Qdrant)
• Building Ranking Classifiers / Learning to Rank (LTR)
  • 📄 Creating great training data
  • 📄 Popular classic models and how they work
  • 📄 Feature engineering
  • 📄 Building and deploying a model end-to-end

Week 3

Nov 18—Nov 21

• Retrieval Augmented Generation (RAG)
  • 📄 Naive RAG
  • 📄 Optimal Chunking Strategies
  • 📄 Agentic & Adaptive RAG
  • 📄 Specialized RAG Techniques (GraphRAG, etc.)
  • 📄 Guardrails
• Interleaving Strategies for RAG (with Max Irwin, Max.io)
• Automating LTR with Click Models and Active Learning
  • 📄 Building click models
  • 📄 Overcoming ranking biases
  • 📄 Active learning
  • 📄 End-to-end Automated LTR

Week 4

Nov 25

• Vector Search Performance Optimization (with Jon Handler – OpenSearch)
  • 📄 ANN Strategies
  • 📄 Representation Learning
  • 📄 Reranking & Recall Optimization
  • 📄 Scalar & Binary Quantization
  • 📄 Product Quantization
  • 📄 Matryoshka Representation Learning
  • 📄 Combining Optimization Strategies
  • 📄 Server selection & Performance Stats at Scale (AWS / OpenSearch)
• Thanksgiving Break (US Holiday)

Week 5

Dec 2—Dec 5

• Semantic Query Understanding
  • 📄 Semantic Query Parsing
  • 📄 “Bag of Documents” approaches
  • 📄 Semantic Knowledge Graphs
  • 📄 Query Classification (index-based)
  • 📄 Query-sense disambiguation
  • 📄 Semantic Caching
  • 📄 Model serving
  • 📄 Wormhole Vectors
• Improving Search Relevance through Semistructured Embeddings (with Daniel Svonava – Superlinked)
• Agentic Search
  • 📄 The Agentic Search paradigm
  • 📄 Query classification (LLM-based)
  • 📄 Query sense disambiguation (LLM-based)
  • 📄 Optimizing search for humans vs. Agents
  • 📄 Search as a series of tool calls for Agent coordinators
  • 📄 Cheating at Search with Agents

How Does the Course Differ From the Book?

One of the top questions we get about the course is how it differs from the AI-Powered Search book. The short answer is there’s about a 30%-40% overlap in content. The course covers a lot of material that was outside the scope of the book, and it also doesn’t cover everything in the book. So the two complement each other very well. Here’s Dmitry Kan asking Trey about this during their recent “Wormhole Vectors” Lightning Talk.

Enrollment Deadline

Signups close on Nov 3, but you can
save 20% if you sign up by 10/30 with code “luckytiming“
save $250 if you sign up by 11/2 with code justintime .

(also, be sure to ask your employer if they have some remaining training budget for 2025 to assist you.)

Don’t miss out on this amazing opportunity to hone your skills for building modern retrieval for humans and agents!