Video course · 9 chapters · 36 min · certificate
AI Agent Design Patterns: I Tested All 35 So You Don't Have To
Know which agent design pattern to reach for: tools and actions, reasoning and reflection, retrieval, memory, sampling and search, multi-agent, safety guardrails and specialty patterns, tested one by one.
What you'll learn
- Explain why design patterns matter for agents
- Use tool use and ReAct loops
- Apply reflection with producer and critic roles
- Choose retrieval and memory patterns
- Use sampling, search and multi-agent patterns well
- Enforce safety with code-level guardrails
Chapters
9 chapters · 35:50-
2:00
01Intro Members
Why patterns matter
There are 35 agentic design patterns; knowing the right one dramatically improves results.
-
4:28
02Tools Members
Tools and actions
Tool use and ReAct: the model reasons, calls a tool, observes and repeats.
-
5:07
03Reflection Members
Reasoning and reflection
Reflection uses a producer that drafts and a critic that grades, then another pass.
-
6:03
04Retrieval Members
Retrieval
Retrieval lets agents answer from your documents, the bridge from chatbot to knowledgeable agent.
-
3:54
05Memory Members
Memory
Memory is state that survives into the next session, turning a stateless model into one that remembers.
-
5:22
06Search Members
Sampling and search
When one answer isn't enough: generate several candidates and pick the best.
-
4:28
07Multi-agent Members
Multi-agent
Supervisors hand subtasks to specialists; popularity has come in waves.
-
2:18
08Safety Members
Safety and guardrails
Let code, not the model, make irreversible calls: hooks, harness logic and dry runs.
-
2:10
09Choosing Members
Specialty patterns and choosing
Specialty patterns like reinforcement-style self-improvement, and which shape to reach for.
Jobs this course suits
Our AI checked this course against 500 jobs; these get the most out of it. Each job links to its learning path.