RAG & Knowledge Systems

This course teaches builders to create and evaluate retrieval-augmented systems that ground AI in their data. It covers why RAG matters, embeddings and vector search, building a knowledge base, and evaluation and pitfalls. Learners finish able to build a RAG system that returns trustworthy answers.

What this course covers

The course runs across 12 topics, each one a short adaptive session rather than a recorded lecture. The tutor explains the idea, works an example, checks understanding, and adjusts the next step to the answer given.

The Case for RAG: Limitations of LLMs and the Power of Grounding — Examine the limitations of parametric memory, including hallucinations and outdated knowledge. Learn how Retrieval-Augmented Generation (RAG) resolves these issues by grounding model outputs in external, verifiable data sources.
Mathematical Foundations of Embeddings — Understand how text is converted into high-dimensional vector representations. Learn about semantic space, distance metrics (Cosine similarity, Euclidean distance, Dot product), and how models capture contextual meaning.
Generating and Manipulating Embeddings in Python — Write code to generate text embeddings using industry-standard API models and local open-source models, comparing vector dimensions and output structures.
Vector Databases and Indexing Strategies — Explore how vector databases store and index high-dimensional vectors for fast retrieval. Learn about indexing algorithms like HNSW (Hierarchical Navigable Small World) and IVF (Inverted File Index).
Document Chunking and Parsing Strategies — Master the art of preparing unstructured data for a knowledge base. Compare fixed-size chunking, recursive character splitting, and semantic chunking to preserve context.
Building a Vector Knowledge Base — Design and implement a complete data pipeline: load source documents, chunk the text, generate embeddings, and upsert them into a vector database.
Querying the Vector Database and Semantic Search — Implement semantic search by embedding user queries, querying the vector database, and retrieving the top-k most relevant document chunks.
Constructing the Contextual Prompt and Generating Answers — Learn how to assemble the retrieved chunks and user query into a secure, structured system prompt that constrains the LLM to answer only using the provided context.
Advanced Retrieval: Hybrid Search and Re-ranking — Improve retrieval accuracy by combining keyword search (BM25) with semantic vector search, and apply cross-encoder re-ranking models to prioritize the best context.
RAG Pitfalls: Lost in the Middle, Noise, and Bad Formatting — Analyze common failure modes in RAG systems, such as context window limits, distracting noise in retrieved documents, and the 'lost in the middle' phenomenon where LLMs ignore middle context.
Quantitative Evaluation: RAG Triad and Metrics — Explore framework-based evaluation using metrics like Faithfulness (hallucination check), Answer Relevance (answering the user's prompt), and Context Precision (retrieval quality).
Implementing an Automated Evaluation Pipeline — Build an end-to-end evaluation script using an evaluation framework (like Ragas or TruLens) to programmatically score your RAG pipeline's performance and identify weak points.

Skills you build

Results are measured against named skills in the OneRange taxonomy of more than 10,000 skills, so a manager sees proficiency per skill rather than a completion tick. This course maps to Vector Databases, Retrieval-Augmented Generation (RAG), Large Language Models (LLMs), Semantic Search, Word Embeddings.

Vector Databases
Retrieval-Augmented Generation (RAG)
Large Language Models (LLMs)
Semantic Search
Word Embeddings

Who it is for

Builders, data & support teams. The material is pitched at advanced level, and takes roughly 180 minutes at a typical pace. Because every session adapts, someone who already knows a topic moves through it quickly instead of sitting through an explanation they do not need.

It sits in the Category: RAG track of the Vero AI Catalog, and can be assigned to one person, a team, or the whole company.

How it is delivered and assessed

Delivery is conversational and interactive, including code lab, guided lab exercises. There is no video to sit through and no slide deck to click past.

Understanding is checked with a code assessment of 5 items, with a pass mark of 70%.

Administrators can copy this course into their own library and adapt it — edit the outline, change the duration, swap the assessment format, or ground it in internal documentation so answers cite the company's own source material.

Also in this track

Embeddings & Vector Databases — A deeper technical layer under RAG: choosing embedding models, operating vector databases, and tuning hybrid search. Learners benchmark retrieval quality and cost across configurations.
Data & Analytics with AI — Query, analyze, and explain data with AI
AI for Customer Experience — Faster, higher-quality support and success

Frequently asked questions

How long does this course take?

It runs roughly 180 minutes at a typical pace. Because every session adapts, someone who already knows a topic moves through it quickly instead of sitting through an explanation they do not need.

How is it assigned to a team?

An administrator assigns it from the Vero dashboard to one person, a team, a department or the whole company, and sees progress and results per person and per skill.

Can we customise it with our own documents and terminology?

Yes. Administrators can copy this course into their own library and adapt it — edit the outline, change the duration, swap the assessment format, or ground it in internal documentation so answers cite the company's own source material.

Course code CAT-RAG.