Why Devin does not use multi agents

Hosted by Jason Liu and Walden Yan

2,299 students

What you'll learn

Multi-agent framework limitations

Students will identify why multi-agent LLM frameworks fail in practice.

Context engineering principles

Students will apply foundational context engineering techniques to improve single-agent LLM performance and reliability.

Alternative architecture strategies

Students will evaluate when to choose single-agent vs multi-agent approaches for different LLM application scenarios.

Why this topic matters

Multi-agent LLM frameworks promise coordinated AI systems but often deliver disappointing results in production. Understanding why they fail—and mastering single-agent context engineering instead—will save you months of wasted development time. You'll learn to build more reliable, maintainable AI applications by choosing the right architectural approach from the start.

You'll learn from

Jason Liu

Consultant at the intersection of Information Retrieval and AI

Jason has built search and recommendation systems for the past 6 years. He has consulted and advised a dozens startups in the last year to improve their RAG systems. He is the creator of the Instructor Python library.

Walden Yan

Co-founder & CPO @ Cognition

worked with

Cognition
Stitch Fix
Meta
University of Waterloo
New York University
© 2025 Maven Learning, Inc.