Building LLM Applications

Building LLM Applications

Duration: 3 Days

Description

This hands-on course teaches developers to build complete, production-minded applications powered by large language models. Building on basic AI development skills, participants learn the architecture of real LLM applications: managing context and conversation state, grounding responses in their own data through retrieval, orchestrating multi-step workflows, and addressing reliability, evaluation, and cost as usage scales. The course also covers current patterns for structured output and tool use, the building blocks that turn an LLM application into an agentic one. Through progressive exercises, attendees build a working LLM application and leave equipped to design and deliver AI features in real systems.
Class time is divided between instructor-led explanation (35 percent) and hands-on development (65 percent).

Audience

Developers and engineers who have built basic AI features and want to create fuller LLM-powered applications. Ideal preparation for agent development and retrieval-focused work, including LangChain/LangGraph and RAG.

Objectives

  • Describe the architecture of a modern LLM application
  • Manage conversation context and application state
  • Ground model responses in custom data using retrieval
  • Orchestrate multi-step LLM workflows
  • Evaluate, test, and improve application quality
  • Address reliability, security, and cost in LLM applications

Prerequisites

Participants should be able to build basic AI-powered features. The following will help:

  • Working knowledge of Python (see Python Foundations)
  • Experience calling AI model APIs (see AI for Developers)
  • Familiarity with generative AI concepts

Related AI Courses

See the full AI training roadmap and course directory for how this fits into a broader learning path.

Continue on to Introduction to AI Agents and LangChain & LangGraph, or pair with RAG Deep Dive to strengthen retrieval-grounded responses.

Course Outline

  • Module 1 – Course Introduction
    • Welcome and objectives
    • Anatomy of an LLM application
    • From single calls to real applications
  • Module 2 – Context and State
    • Managing conversation history
    • Context windows and their limits
    • Application state and memory strategies
    • Hands-on exercise: a stateful chat feature
  • Module 3 – Grounding with Retrieval
    • Why grounding matters: reducing hallucination
    • Introduction to retrieval-augmented generation
    • Embeddings and semantic search
    • Connecting an LLM to your own data
  • Module 4 – Orchestrating Workflows
    • Chaining multiple LLM calls
    • Routing, branching, and tool use
    • Structured pipelines
    • Hands-on exercise: a multi-step workflow
  • Module 5 – Quality and Evaluation
    • Defining quality for LLM output
    • Evaluation strategies and test sets
    • Prompt and pipeline iteration
    • Monitoring in production
  • Module 6 – Reliability, Security, and Cost
    • Error handling and graceful degradation
    • Input validation and prompt-injection awareness
    • Managing token usage and cost
    • Performance and scaling considerations
  • Module 7 – Capstone Build
    • Designing a complete LLM application
    • Integrating retrieval, workflow, and evaluation
    • Hands-on exercise: build an end-to-end application
  • Module 8 – Course Wrap-Up
    • Key takeaways
    • Resources for continued learning
    • Next steps toward agents