Advanced Text Generation Techniques

In this lab, you will explore advanced methods for grounding text generation in facts and automating multi-step reasoning with agentic workflows. You’ll learn how retrieval-augmented generation (RAG) systems work, and how to implement a lightweight RAG pipeline with an agentic workflow.

Learning objectives

Upon completion of this lab, you will be able to:

• Describe and implement agentic workflows for multi-step reasoning and task automation.
• Discuss memory management and conversation handling in LLM applications.
• Implement the architecture and components of RAG pipelines, including vector stores and retrievers.
• Compare the pros and cons of agentic techniques in real-world scenarios.

Intended audience

This course is designed for:

• Data Scientists
• Software Developers
• Machine Learning Engineers
• AI Engineers
• DevOps Engineers

Prerequisites

Completion of previous modules is highly recommended before attempting this lab.

Lab structure

Demo: RAG + Agentic Workflow — Study Abroad Advisor
In this demo, you will build a lightweight retrieval-augmented agentic workflow:
- Load a small knowledge base of study abroad handbooks (Visa & Compliance, Housing & Health, Travel & Insurance).
- Chunk and embed documents, and build a FAISS retriever.
- Wire an agentic graph with a “retrieve → generate” pattern.
- Query the system with student-style questions and receive concise, cited answers grounded in the documents.

Intended learning outcomes:
- Implement a minimal RAG pipeline with citations.
- Control chunking and retrieval parameters to balance precision and recall.
- Use an agentic workflow to structure retrieval and generation.