Agentic AI Engineering
Beyond single-shot LLM calls lies agentic AI — systems that reason, call tools, maintain state and execute multi-step workflows autonomously. This module builds and evaluates real AI agents designed to operate safely in production.
What You Will Learn
A detailed, industry-aligned breakdown of every topic covered in this module.
- AI agent fundamentals and architecture
- Tool calling and function calling
- State and memory management
- Routing and agent workflows
- Model Context Protocol (MCP)
- Multi-agent patterns
- Human-in-the-loop design
- Retries, fallbacks and guardrails
- Agent evaluation and observability
- Prompt injection defence
Tools You Will Use
Hands-on time with the same tools used in professional AI engineering and production ML workflows.
Python
Core programming language used across every module, from scripting to AI model development.
LangGraph
Agent orchestration framework used to build stateful, multi-step AI agent workflows.
MCP
Model Context Protocol used to connect AI agents and LLMs to external tools and data sources.
LLM APIs
Large language model APIs used to integrate generative AI capabilities into applications.
FastAPI
Modern Python web framework used to build production-grade backend and AI-serving APIs.
PostgreSQL
Relational database used to store structured application and AI system data.
Vector DBs
Purpose-built databases used to store and retrieve embeddings for RAG and semantic search.
Docker
Containerization platform used to package and deploy applications and AI services consistently.
Hands-On Labs
Production-style AI engineering lab scenarios, built using the same stack real AI teams ship with.
Design and build an AI agent with tool-calling and memory management using LangGraph.
Implement multi-step agent workflows with routing and human-in-the-loop checkpoints.
Connect an agent to external tools and data sources using the Model Context Protocol.
Add retries, fallbacks and guardrails to make an agent reliable in production.
Evaluate agent performance and add observability for monitoring agent behaviour.
Assessment
Knowledge Assessment
Quiz covering agent architecture, tool calling, MCP and agent safety/guardrail concepts.
Practical Evaluation
Students must design, build and evaluate a multi-step AI agent that safely completes a real task using external tools.
Projects
Industry-style deliverables added directly to your project portfolio.
Agentic AI Application
Build intelligent agents that use tools, make decisions and execute multi-step tasks.
What This Module Builds
Students design, build and evaluate AI agents that call tools, maintain state, execute multi-step workflows, and operate safely and reliably in production.