💻 AI Agents Fix Flaky Tests at Kong
Kong engineers have implemented an agentic AI workflow to automate the fixing of flaky tests in their CI/CD environment. The system utilizes a multi-agent architecture based on Claude 3 Opus and Claude 3 Haiku. The agent successfully fixed 12 out of 15 tests, including a bug that humans had been unable to resolve for five years.
🌍 Moving to multi-agent systems allows for the automation of complex debugging cycles, reducing token costs.
👤 Automating technical debt management frees up engineers' time for new tasks.
Source 1: https://konghq.com/blog/engineering/how-we-used-agentic-ai-to-fix-kong-gateways-flakiest-tests