RAG and AI agents: assistants built on company data

Design, evaluate and ship assistants that answer from your documents and agents that use your systems.

  • Audience:Technical
  • Duration:16 hours · 2 days
  • Format:On-site or online
  • Language:Italian or English

What you'll learn

  • Design a RAG pipeline: chunking, embeddings, retrieval and answer
  • Measure retrieval and answer quality with repeatable tests
  • Give an agent tools and permissions, with handover to a person when needed
  • Manage cost, latency and knowledge base updates
  • Monitor an assistant in production

Course content

4 modules · 16 hours in total

1. RAG architecture4 hours
  • Chunking by section, not by fixed length
  • Embeddings and vector stores
  • Hybrid search and re-ranking
2. Evaluation4 hours
  • Reference question sets
  • Retrieval and answer metrics
  • Relevance thresholds: when it is better not to answer
3. AI agents4 hours
  • Tool calling and integration with APIs and CRMs
  • Least privilege and human confirmation
  • Handover to an operator with a case summary
4. In production4 hours
  • Logging, traces and conversation analysis
  • Cost per request and caching
  • Lab: an assistant on your documents

Requirements

  • Basic Python
  • Familiarity with REST APIs
  • Ideally the "How language models work" course

Description

An assistant that answers from company documents looks simple in a demo. In production the problems are different: confident but wrong answers, documents split in the wrong place, growing costs, agents with too many permissions.

These are the same problems we solve in our projects, from customer support to internal knowledge bases. Over two days you build an assistant on your own documents and learn to measure it before putting it in front of a user.

Who this course is for

  • Developers and data engineers
  • Technical teams building an internal assistant or agent
  • Technical leads assessing vendors and architectures

Request information

The course is not sold online: we run it for your team. Write to us and we will send you the programme and a proposal.

Who is asking? *

We reply within 24 business hours with the programme and a proposal. No commitment.

More courses in this category