About the course
This hands-on instructor-led workshop provides a deep dive into using Elasticsearch as a high-performance vector database and engine for Retrieval-Augmented Generation (RAG).
You will learn how to transform unstructured data into searchable vectors and integrate them into a production-ready RAG pipeline that grounds Large Language Models (LLMs) in private, domain-specific data.
The curriculum focuses on the latest stable features of the Elastic Stack, including the Inference API, the ELSER (Elastic Learned Sparse EncodeR) model, and hybrid search techniques.
By the end of the session, attendees will be able to build a retrieval system that combines the precision of keyword matching with the semantic depth of vector embeddings to eliminate LLM hallucinations and provide factually grounded responses.
For private courses where the audience already has strong experience of Elastic Stack, we are happy to tailor the duration and modules according to needs (for instance condensing the workshop to a 1-day intensive high-level exploration of capabilities).
Instructor-led online and in-house face-to-face options are available - as part of a wider customised training programme, or as a standalone workshop, on-site at your offices or at one of many flexible meeting spaces in the UK and around the World.
-
By the end of this course, attendees will be able to:
- Configure Elasticsearch as a vector store using Hierarchical Navigable Small World (HNSW) indexing.
- Implement the Elastic Inference API to manage text embeddings without external orchestration tools.
- Develop hybrid search queries that merge BM25 scores with k-nearest neighbor (kNN) vector results.
- Design efficient document chunking and metadata filtering strategies to improve retrieval precision.
- Construct a complete RAG workflow that connects Elasticsearch retrieval to LLM generative outputs.
-
This course is designed for Database Administrators, Data Engineers, and Software Architects who are tasked with implementing AI-powered search features. It is particularly suited for professionals looking to transition traditional search or relational database expertise into the field of Generative AI and vector search architecture.
-
Foundational knowledge of database concepts such as indexing, schemas, and CRUD operations.
Familiarity with Python for basic scripting and API interactions.
A basic understanding of JSON-based REST APIs.
No prior experience with Elasticsearch or Vector Databases is required.
We can customise the training to match your team's experience and needs - for instance with more time and coverage of fundamentals for budding new data professionals. Get in touch to find out more.
-
This RAG apps with Elasticsearch course is available for private / custom delivery for your team - as an in-house face-to-face workshop at your location of choice, or as online instructor-led training via MS Teams (or your own preferred platform).
Get in touch to find out how we can deliver tailored training which focuses on your project requirements and learning goals.
-
Fundamentals of Vector Search in Elastic
The transition from inverted indices to dense vector representations.
Understanding the Dense Vector field type and element sizes.
Configuring HNSW (Hierarchical Navigable Small World) for low-latency similarity searches.
Introduction to Similarity Metrics: Cosine Similarity, Dot Product, and L2 Norm.
Data Ingestion and the Inference API
Setting up Ingest Pipelines for automated data transformation.
Using the Elastic Inference API to generate embeddings within the cluster.
Managing text-embedding models and the ELSER sparse encoder.
Document chunking strategies: Fixed-size, overlapping, and semantic boundary methods.
Advanced Retrieval Strategies for RAG
Executing kNN (k-Nearest Neighbor) searches via the Search API.
Implementing Hybrid Search using Reciprocal Rank Fusion (RRF).
Applying Metadata Filtering to restrict vector search space.
Tuning BM25 keyword matching alongside semantic results for maximum precision.
LLM Integration and RAG Orchestration
Context Window management: Selecting the most relevant "top-k" results.
Prompt Engineering: Designing templates that ground LLMs in retrieved context.
Handling "No Result" scenarios and hallucination mitigation.
Evaluating retrieval quality using precision and recall metrics.
Trusted by
