Build retrieval systems you can measure.
Move from keyword search to dense and hybrid retrieval, chunking, reranking, HyDE, ANN indexes, and rigorous ranking evaluation.
Build lexical, dense, and hybrid retrieval
Choose chunking and diversification strategies
Add reranking and query transformation
Evaluate retrieval with Recall@k, MRR, and nDCG
Helpwell, a customer-support knowledge corpus whose retrieval quality is measured before and after every technique.
Basic Python and a high-level understanding of LLM applications are helpful.
Open any module to review its exact sections. Progress and completion follow you through the course.
Topics include RAG vs fine-tune vs long-context, Lost in the middle, Chunking, and more.
Topics include Vector space, Cosine vs dot, Normalization, and more.
Topics include Fixed/recursive/semantic chunking, Overlap, Chunk size, and more.
Topics include When dense fails, Reciprocal Rank Fusion, Weighted hybrid, and more.
Topics include Brute force cost, IVF / k-means, HNSW graph walk, and more.
Topics include Bi- vs cross-encoder, Two-stage retrieval, Recall@k, and more.
Topics include HyDE, Multi-query, Step-back, and more.
Topics include Context assembly, Grounding, Citations, and more.
AI engineers building RAG applications
Search and data practitioners
Technical teams evaluating vector databases
These are independent LDS courses that extend the closest skills. Choose the direction that matches what you want to build next.
Engineer what the model sees and returns.
Replace “it feels better” with evidence.
Turn one LLM call into a dependable service.
The first module establishes the language and example used throughout the rest of the course.
