AI FDE: The Forward Deployed Engineer & Agentic AI Architect (Udemy)

Our take

AI FDE: The Forward Deployed Engineer & Agentic AI Architect is a all level AI Agents & Agentic AI course from Ed Donner on Udemy. It runs about 21.5 hours, and holds 4.9★ from 447 students in our Udemy course reviews.

Best for: all level learners who want one guided path through AI Agents & Agentic AI. Skip it if: You already work professionally in AI Agents & Agentic AI and need advanced architecture patterns.

📘 What you will learn

• Identify the skills of an AI Architect and Forward Deployed Engineer (FDE) with the roadmap to get there
• Accelerate delivery with Claude Code and the Agentic Development Lifecycle (ADLC) while staying accountable
• Apply Harness Engineering to build agents with LangChain, Deep Agents and the Claude Agent SDK
• Choose between frontier, open-source and local LLMs on intelligence, speed, token cost and privacy
• Use evals and observability in Langfuse to drive agent architecture decisions and measure business outcomes
• Give agents long-term memory with Agentic RAG over MCP using vector datastores
• Deliver production-grade agent workflows in LangGraph with durable state, crash recovery and human in the loop
• Scale and optimize with agent swarms, LLM as a judge and recursive self-improvement
• Lead AI initiatives with your clients from discovery and metrics to governance, responsible AI and handover
• Secure agents against prompt injection and the lethal trifecta with guardrails, policies and AI gateways
• Create a data and integration architecture for agents with an enterprise context layer and MCP gateway
• Deploy LangGraph agents to production with AWS Bedrock AgentCore Runtime, Gateway, Policy and Memory

📑 Course content

3 sections · 124 lectures · 21.5 hours of video · updated 10/2026

1. Getting started — 45 lectures · 7.8h
2. Core foundations — 35 lectures · 6.1h
3. Building real projects — 44 lectures · 7.6h

🎯 Who it is for

Suits: Complete beginners, career switchers, and self-taught devs who want structure.

Needs: Some experience with Python and LLMs is ideal but not required

👤 Instructor

Ed Donner — Ed Donner is a coder, speaker and leader in the field of Agentic AI, Forward Deployed Engineering and LLMs. He’s the co-founder and CTO of AI startup Nebula, a former Managing Director at JPMorgan Chase, and a well-regarded lecturer on LLMs and Agents with over a million enrollments across 194 countries. He previously founded AI startup untapt, acquired in 2021, after a 15 year career in technology on Wall street. He holds a patent in a deep learning matching model, and an MA in Physics from Oxford. When he’s not at work experimenting with LLMs, Ed can usually be found at home.. experimenting with LLMs.

❓ FAQ

No written reviews yet. Be the first to review this course →