Agentic AI Engineering (Build, don't just prompt)
Agentic AI Engineering (Build, don't just prompt)
Move beyond basic prompting and enter the world of Agentic AI Engineering. This course is designed for the 2026 market, where the demand is for engineers who can build autonomous, self-correcting agents using Python, LangGraph, and CrewAI. You will learn to architect complex agentic workflows, manage state, deploy via CLI using Docker, and master the '50 lines of code' challenge. From low-level API management to multi-agent orchestration, this is the definitive guide to building the future of autonomous software.
Lessons
- Introduction to Agentic AI: Beyond the Prompt
- Setting Up Your Engineering Environment
- Dockerizing the Execution Environment
- The LLM as a Reasoning Engine (API basics)
- Introduction to Function Calling
- Project: Simple Weather Agent in < 50 Lines
- Agentic Frameworks Overview: LangGraph vs CrewAI
- LangGraph Core: Nodes and Edges
- Designing a State Machine for your Agent
- Conditional Edges: Letting the Agent Decide
- CrewAI: Role-Based Agent Design
- CrewAI: Tasks and Delegation
- Building a CLI for your Agents
- Managing API Rate Limits and Retries
- 50-Line Challenge: Research Agent
- Advanced LangGraph: Persistence and Checkpointing
- Tool Integration: Custom Python Tools
- Streaming Agent Outputs in Real-time
- Handling Sensitive Data & Env Secrets
- 50-Line Challenge: Auto-Code Fixer
- Advanced CrewAI: Custom Processes
- Agent Memory: RAG Integration
- Testing Agents: Evaluating Accuracy
- Deploying Agents to the Cloud (Render/Railway)
- 50-Line Challenge: Social Media Manager