Retrieval-Augmented Generation and Vector Databases

Build retrieval-augmented generation systems

Large language models can generate powerful responses, but they’re limited by the information they were trained on. Retrieval augmented generation (RAG) helps overcome this by connecting LLMs to external sources of knowledge.

On this three-week course, you’ll learn how RAG systems work and develop practical skills to build, customise, and evaluate them. You’ll explore text embeddings, semantic similarity, and vector search before working with vector databases including Chroma, Pinecone, and Weaviate.

Work with vector databases and semantic search

Discover how vector databases store and retrieve information based on semantic meaning rather than simple keyword matching. You’ll learn how to create indexes, manage metadata, perform CRUD operations, and implement similarity searches.

You’ll also explore different embedding models and vector database technologies. By comparing different approaches, you’ll develop the knowledge to select the right tools for your RAG applications.

Build and evaluate advanced RAG applications

Learn how to build complete RAG pipelines using retriever, augmenter, and generator components. You’ll work with document loaders and retrieval chains, customise prompts, and explore advanced techniques such as re-ranking, query transformation, hybrid search, and self-querying retrievers.

You’ll also learn how to evaluate retrieval and generation quality, diagnose weaknesses, and improve your pipelines. The course culminates in a hands-on project where you’ll build a document-based knowledge assistant capable of producing accurate, grounded answers.

By the end, you’ll be able to design, build, and evaluate RAG systems, preparing you for roles involving LLM application development and retrieval-augmented generation.

This course is designed for learners who can build LLM applications with APIs and frameworks and who now want to build accurate, knowledge-grounded systems.

It is ideal for developers building retrieval-based assistants, data professionals creating document Q&A tools, and engineers preparing for RAG-focused LLM roles. Learners should be comfortable with LLM APIs and LangChain or LlamaIndex before starting. By the end, they will be able to design, build, and evaluate production-quality RAG systems.

This course is designed for learners who can build LLM applications with APIs and frameworks and who now want to build accurate, knowledge-grounded systems.

It is ideal for developers building retrieval-based assistants, data professionals creating document Q&A tools, and engineers preparing for RAG-focused LLM roles. Learners should be comfortable with LLM APIs and LangChain or LlamaIndex before starting. By the end, they will be able to design, build, and evaluate production-quality RAG systems.

• Explain how embeddings, vector databases, semantic similarity, and Retrieval-Augmented Generation support grounded LLM responses.
• Apply document loading, chunking, embedding, indexing, and retrieval techniques to build a functional RAG pipeline.
• Assess retrieval performance using re-ranking, query transformation, hybrid search, and specialised retriever strategies.
• Evaluate RAG quality using retrieval metrics, faithfulness, relevance, answer correctness, and RAGAS-based assessment.
• Create an end-to-end document-based knowledge assistant that delivers accurate, grounded, and context-aware responses.