Agentic Development

Who This Is For

• Developers building agentic features, not just prototypes
• Product engineers integrating AI agents into workflows
• Tech leads who need dependable patterns for autonomous systems
• Founders building AI-native products who need speed + rigor

What You’ll Build

By the end of this course, you will have a production-ready agentic workflow — complete with tools, memory, evaluation harnesses, and governance checks. You’ll know how to decide when agents are appropriate, how to keep them safe, and how to ship them with confidence.

Learning Objectives

• Define when an agentic architecture is the right choice versus simpler automation
• Build agent contracts, tool interfaces, and memory policies that reduce drift
• Design single-agent and multi-agent runtimes with clear stop conditions
• Add evaluation harnesses, observability, and safety controls for production reliability
• Ship a complete agentic workflow with governance, cost, and latency constraints

Course Structure

Lesson 1: Agentic Development: What It Is (and What It Isn’t)
Lesson 2: Agent Runtime Basics: Models, Loops, and State
Lesson 3: Agent Contracts: Roles, Prompts, and Guardrails
Lesson 4: Tooling & Action Interfaces
Lesson 5: Memory Systems: Short-Term, Long-Term, Retrieval
Lesson 6: Planning & Decomposition Patterns
Lesson 7: Multi-Agent Orchestration
Lesson 8: Human-in-the-Loop Controls
Lesson 9: Evaluation & Testing for Agents
Lesson 10: Reliability, Observability, and Safety
Lesson 11: Cost, Latency, and Scaling
Lesson 12: Shipping Agentic Products

Each lesson includes knowledge checks, recall cards, hands-on exercises, and real-world scenarios so you can apply the patterns immediately.

Mini Projects

1. Agent Triage Bot (Lessons 3-4): Build a support triage agent with a strict tool schema, priority classification, and escalation policy.
2. Research -> Brief Pipeline (Lessons 6-7): Orchestrate two agents where one collects source material and another produces a structured brief with citations.
3. Agent Regression Harness (Lesson 9): Create a repeatable 10-15 case evaluation suite with scoring, failure analysis, and regression checks.

Capstone Project

Agentic Product Sprint: Design and ship an end-to-end agentic workflow that drives a real business outcome. You will define roles, tool contracts, memory rules, evaluation criteria, and a production readiness checklist before launch.

• All courses included
• New courses added automatically
• Manage or cancel anytime

Understand the building blocks of an agent runtime and the minimum viable loop