INTERMEDIATE · 8 MODULES · 64 LESSONS
Multi-Agent Orchestration
Decide when it's worth splitting work across several agents and, when it is, coordinate them by hand in Python without relying on any framework.
- For you if
- You've already built an agent with tools and some tasks are too much for it alone.
- By the end
- You'll know how to apply five coordination patterns, measure what coordination costs, and pick the right pattern for each task.
- You need
- Python 3.14 with some dataclasses and concurrency. No API key, internet, or frameworks needed.

Curriculum
04Parallel Fan Out And Aggregation
- Module 4: Parallel fan-out and aggregation
- Identifying independent sub-tasks
- The base case: fixed, enumerated order
- Aggregating results deterministically
- Fan-out inside a single agent: the pricing_agent variant
- Real concurrency with ThreadPoolExecutor
- Measuring the latency savings in rounds
- Mini-project: fan-out across Reservo
05Handoff And Delegation
- Module 5: Handoff and Delegation
- Recognizing When a Handoff Is Needed
- The Handoff Package: What Goes In, What Stays Out
- The Handoff Tool and the Interrupted Loop
- Reservo's Handoff, Executed End to End
- Guarding Against Endless Handoff Chains
- Measuring the Cost of a Direct Handoff
- Mini-Project: Handoffs in Reservo
06Shared State And The Blackboard Pattern
07Orchestrating The Full Reservo System
- Module 7 Introduction: Combining the Patterns Over a Real Request
- Anatomy of a Compound Request
- The Supervisor Decides, the Pipeline Validates Before Booking
- The Independent Questions Run in Fan-out While the Pipeline Resolves
- An Agent Hands Off Mid-way Through a Fan-out Branch
- A Blackboard, Shared by All Three Patterns at Once
- The Full Composed Run, Executed End to End
- Mini-Project: Three Compound Requests for Reservo
08Project The Reservo Multi Agent System
Reviews
What students say
These reviews are from enrolled students who completed at least 50% of the course. We moderate reviews only on content grounds (spam, offensive language, personal data), never for being critical or negative.
No approved reviews yet.
Be the first to share your experience!