From 0 to your first RAG prototype
Three days in which you don't just understand RAG but build it yourself: evolving from the naive pipeline to a modular architecture, every step motivated by a concrete problem and quantitatively validated.
Between a RAG prototype and a production system lies a field of technical decisions that influence one another. Generic defaults don't work. This workshop makes the difference transparent and measurable, until you master it yourself.
01 - THE STRUCTURE
Three days, one common thread
From the first working prototype to advanced architecture, each day builds on the previous one.
Day 1
Naive RAG: foundation and its limits
You build your first working RAG pipeline and understand systematically why it isn't yet enough for production.
- The full RAG path: from question to sourced answer
- The building blocks of a pipeline and how they interact
- A first working pipeline on the practice dataset
- Failure modes: where naive RAG systematically breaks down
Day 2
Retrieval quality: the first upgrades
Every upgrade is motivated by a concrete problem and quantitatively validated on a shared test set.
- Hybrid search (BM25 + vector) and reranking
- Metadata filtering and chunking strategies
- Contextual retrieval
- Quantitative evaluation with Ragas
Day 3
Going deeper & advanced architectures
You learn to weigh the levers that determine quality, cost, and operations, and assess future RAG developments on your own.
- Embeddings: selection and impact
- The vector-DB landscape compared
- Advanced RAG architectures
- Production operations: running, monitoring, scaling
02 - THE METHOD
Why this isn't a webinar
Engineering instead of slides: you work on a real pipeline and measure every bit of progress.
Evolutionary, not theory
A step-by-step build from naive to modular RAG. You see each architecture emerge instead of consuming finished answers.
Problem-driven
Every upgrade is motivated by a concrete problem, no feature without a reason.
Quantitatively validated
Every improvement is measured on a shared test set (Ragas). You learn to prove, not to guess.
Hands-on
Throughout on a realistic practice dataset, transferable to your own documents.
03 - INDEPENDENCE
Which tool is right, you decide for yourself.
You learn to make well-founded decisions: build vs. buy and which stack fits which case. We weigh the Microsoft ecosystem, Copilot, Azure AI Search, Fabric, just as neutrally as open-source and best-of-breed stacks like Vespa.ai, Qdrant, pgvector, and Chroma.
By the end, you can judge for yourself what your project needs, from a ready-made platform to a custom build.
04 - YOUR OUTCOME
What you take away
After three days you can independently design, implement, and evaluate a production-ready RAG pipeline, and assess future RAG developments yourself.
- A RAG pipeline you built and evaluated yourself
- A decision framework for build vs. buy and stack selection
- A quantitative evaluation methodology (Ragas) for your own projects
- The ability to classify new RAG techniques independently
Production-ready RAG expertise in your team
05 - WHO IT'S FOR
Who is the workshop for?
Technical leads & architects
Plan RAG pipelines or make existing prototypes production-ready.
Product owners
Understand which levers influence answer quality.
IT decision-makers & CTOs
Make build-vs-buy and cloud-vs-on-prem trade-offs with confidence and plan budgets.
06 - THE WORKSHOP LADDER
Two ways to learn RAG
Start self-guided or go deep in the guided 3-day workshop.
Self-Guided RAG Workshop
A free entry point to work through yourself.
RAG Workshop · 3 days
The guided deep dive from naive to modular RAG.
07 - FAQ
Frequently asked questions
What are the prerequisites for the workshop?
Basic programming skills in the team help. We provide a practice dataset, and you can bring your own documents.
Does the workshop run on-site or remote?
Both are possible, on-site or remote, depending on what works better for your team.
How many people can take part?
Up to 10 participants, so there is enough room for questions and hands-on work.
What is the concrete result?
A RAG pipeline you built and evaluated yourself, a quantitative evaluation methodology with Ragas, and a decision framework for build vs. buy and stack selection.
Do we need a finished RAG idea beforehand?
No. We start from the basics and work up to a production-ready architecture step by step. A concrete idea helps but is not required.
Is the workshop tied to a vendor?
No. We weigh Microsoft, open source, and best-of-breed stacks neutrally. You learn to make the right choice for your case yourself.
Self-Guided RAG Workshop: your RAG architecture step by step
Work through the key architecture decisions for your RAG pipeline at your own pace: structured, hands-on and ready to apply.
- Chunking strategy for your document types
- Retrieval architecture: vector, hybrid or knowledge graph?
- Reranking: improving the quality of results
- Model selection: LLM, embedding and reranker
- Operating model: cloud, on-premise or hybrid?
Request the workshop
Enter your email address and we will send you the Self-Guided RAG Workshop.
Ready to master RAG yourself?
Book your free 30-minute consultation. We'll clarify your starting point and give you an honest assessment.
Let’s talk about how AI can move your business forward.
Get in touch, we look forward to your project.