Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
May 29, 2025
·
Medellín
LangGraph Multi-Agent Evaluation
This talk demonstrates using agentic methods to autonomously test LangGraph multi-agent systems, identifying vulnerabilities and generating actionable improvement reports.
Overview
In this demo, we showcase how an agentic approach can be used to test and identify vulnerabilities in LangGraph-based agent systems. The system autonomously explores the target LangGraph structure, generates specialized tester agents, creates and executes test cases, invokes the target graph, and produces detailed reports with suggestions for improvement—demonstrating how agents can evaluate and enhance other agents.
Links
Demonstrates LangGraph implementation for inspecting complex Generative AI agent system workflows.
Tech stack