New VDF AI Enterprise is now available on AWS Marketplace
Free learning path 3

Private RAG Engineering

This free course teaches private RAG engineering: building retrieval-augmented generation that runs inside your own infrastructure. You connect a real source, index it, retrieve from databases and tables, keep retrieval inside the source’s permissions and test retrieval quality before rollout, all hands-on in VDF AI Data with an on-premises embedding model.

  • 4 lessons
  • 1 h 45 min
  • Beginner
  • Free, self-paced
Start lesson 1 → Open VDF AI free
Outcomes

What you will be able to do

  • Build a retrieval pipeline from a source system to ranked, cited passages
  • Index documents with settings that suit the content you actually have
  • Answer questions over databases and tables, not just prose
  • Keep retrieval inside the permissions of its source
  • Measure retrieval quality before users see the results
Before you start

Who it is for and what you need

Data engineers, knowledge managers and AI engineers responsible for enterprise search and assistants.

  • Access to a VDF AI deployment with VDF AI Data enabled
  • A handful of documents or a test database you are allowed to index
The lessons

4 lessons, in order

Each lesson is a hands-on build in your own workspace. For every field and option, see the VDF AI Data documentation.

  1. 1 How to Build a Private RAG Pipeline, Step by Step Build a private RAG retrieval pipeline in VDF AI Data. Connect a source with a read-only account, check the data, embed it on-premises and test semantic search. Video
    25 min
  2. 2 How to Build RAG over SQL Databases and Tables Make database rows retrievable for RAG. See which columns VDF AI embeds, why a raw table finds the problem but not the fix, and build a table that answers. Tutorial
    25 min
  3. 3 How to Build Permission-Aware RAG with Access Control Keep RAG inside existing permissions. Enforce access at the source account, keep sensitive rows out of the index, and test who can search it in VDF AI Data. Tutorial
    25 min
  4. 4 How to Test RAG Retrieval Quality Before Rollout Test RAG retrieval before users rely on it. Write hard questions, check the passage holds the answer, compare chunk sizes and keep what measures better. Tutorial
    30 min

Take it to production

Production RAG over regulated data needs the whole pipeline inside your perimeter: sources, indexes, models and the audit trail.

Book an architecture call Enterprise RAG
FAQ

Frequently asked questions

Does any data leave our network in this path?

Not in our deployment: the source database stays where it is, and both the articles and the questions are embedded by bge-m3, a model served on-premises. Check which embedding model your own deployment uses before you index anything sensitive.

Can I follow the path with my own database?

Yes. The lessons use a small synthetic PostgreSQL service desk database, but any source VDF AI Data connects to will do, including Microsoft SQL Server, MySQL and Oracle. Give the connection a read-only account.

Does this path cover generating the answer?

It concentrates on retrieval, because an answer can only be as good as the passages a model is given. Each lesson shows what the retrieved passages and their source IDs are used for next.

After this path

Keep going

Try another free path, or the course that follows this one. It is free for VDF AI customers and partners.

Instructor-led · Advanced

Enterprise RAG Engineering

Engineer governed RAG over documents and databases.

4 live half-days 6 modules + capstone VDF AI Certified RAG Engineer
Free path

VDF AI Fundamentals

Start here: the platform, chat and answers you can check.

3 lessons 1 h
Free path

Self-Hosting Basics

Plan and run self-hosted VDF AI, from portal to runbook.

4 lessons 36 min