Designing Multi-Agent AI Systems

Aki Wijesundara, PhD

AI Founder | Google AI Accelerator Alum

Manu Jayawardana

Exited AI Founder | Co-Founder, Snapdrum

Build multi-agent AI systems that actually work, not just demo

Many aspiring AI builders get stuck at the start. They know what multi-agent AI can do but struggle to turn ideas into systems that actually work in production. Without a clear design approach, agents fail, coordination breaks down, costs blow up, and hallucinations derail results.

This course guides you from zero to production-ready multi-agent AI thinking. Over the course, you will learn to:

  • Design agent roles and responsibilities from scratch even with no prior experience

  • Manage context, memory, and coordination so agents reason reliably

  • Plan, execute, and integrate tools safely within agent workflows

  • Evaluate, debug, and improve agent behavior iteratively

By the end, you will not just understand the theory. You will build functioning multi-agent systems, know how agents coordinate and solve real problems, and gain the confidence to ship reliable AI products, setting the foundation for advanced AI system design and production deployment.

What you’ll learn

Build, manage, and scale data pipelines from scratch — even if you’ve never coded one before.

  • Learn what agents are and how they work together

  • Understand how multi-agent systems differ from single-agent workflows

  • Get a clear mental Get a clear mental model of agent roles, responsibilities, and coordination

  • Define agent roles and tasks from the ground up

  • Learn how to manage memory, context, and interactions so agents reason reliably

  • Integrate tools safely and make agents perform real actions

  • Detect and fix problems early before they affect the whole system

  • Iterate on your system to make it reliable and scalable

  • Walk away with a fully functioning multi-agent AI system you built yourself

Learn directly from Aki & Manu

Aki Wijesundara, PhD

Aki Wijesundara, PhD

AI Founder | Educator | Google AI Accelerator Alum

Google
Meta
OpenAI
Amazon Web Services
NVIDIA
Manu Jayawardana

Manu Jayawardana

Exited AI Founder (Rise AI: 35k Users) | Co-Founder of Krybe and Snapdrum.com

Previous Students from
Google
McKinsey & Company
Boston Consulting Group (BCG)
NVIDIA
OpenAI

Who this course is for

  • Engineers building AI features or platforms

  • Product managers owning AI initiatives

  • Founders designing AI-powered products

What's included

Live sessions

Learn directly from Aki Wijesundara, PhD & Manu Jayawardana in a real-time, interactive format.

Project-based learning

Work on practical exercises that simulate real-world data workflows from ingestion to analytics-ready tables.

Downloadable resources

Access code templates, cheat sheets, and pipeline examples to practice and reuse after the course.

Community of learners

Collaborate, ask questions, and share insights with peers in a supportive learning environment.

Certificate of completion

Showcase your data engineering skills to employers, clients, or your team.

Guided workflow playbooks

Step-by-step guides for building reliable pipelines using best practices and beginner-friendly tools.

Maven Guarantee

This course is backed by the Maven Guarantee. Students are eligible for a full refund up until the halfway point of the course.

Course syllabus

4 live sessions • 118 lessons

Week 1

Feb 1

    Feb

    1

    Foundations of Multi-Agent AI

    Sun 2/16:00 PM—8:00 PM (UTC)

    What is an agent and what is a multi-agent system

    5 items

    Single-agent vs multi-agent workflows

    5 items

    Real-world examples of multi-agent AI

    5 items

    Core coordination patterns and communication methods

    5 items

    Common challenges in multi-agent systems

    5 items

    Hands-On / Outcome:

    3 items

    Assignment To Do

    0 items

Week 2

Feb 2—Feb 8

    Feb

    8

    Designing Agents and Coordination

    Sun 2/85:30 PM—7:30 PM (UTC)

    Defining agent roles: executor, planner, reviewer, supervisor

    6 items

    Delegation and handoffs between agents

    5 items

    Preventing coordination breakdowns

    5 items

    Handling context and short-term memory

    5 items

    Safe planning and basic decision-making logic

    5 items

    Hands-On / Outcome:

    3 items

    Assignment To Do

    0 items

Schedule

Live sessions

2 hrs / week

You will learn core concepts into multi-agents

    • Sun, Feb 1

      6:00 PM—8:00 PM (UTC)

    • Sun, Feb 8

      5:30 PM—7:30 PM (UTC)

    • Sun, Feb 15

      5:30 PM—7:30 PM (UTC)

    • Sun, Feb 22

      5:30 PM—7:30 PM (UTC)

Hands On Projects

4 hrs / week

Complete practical exercises and mini-projects that simulate real-world multi-agent AI workflows. Apply what you learn in live sessions to design agents, coordinate tasks, integrate tools, and manage memory, ensuring you gain hands-on experience that prepares you to build production-ready multi-agent systems.

$999

USD

2 days left to enroll

Enroll