Uploaded June 2026 | Updated September 2026, 3 weeks ago
What happens when one of the world’s leading AI coding companies turns its own product inward?
In this session, Cursor shares how its engineering teams use AI agents, automated evaluations, and agent-powered workflows to build Cursor itself. Learn how the company has evolved from AI-assisted coding to systems of specialized agents that plan, implement, review, test, and improve code with increasing autonomy.
The talk explores how Cursor structures its internal “agent factory,” including custom skills, automated code review agents, event-driven workflows, long-running optimization agents, and evaluation systems that enable scalable software development. You’ll also learn practical approaches to agent verification, CI/CD automation, risk-based code review, and continuous improvement loops that help agents learn from production feedback.
Key Takeaways
• AI coding is evolving from assistants to systems of specialized agents working together.
• Cursor uses custom skills, automations, and evaluation systems to improve agent performance over time.
• Verification is critical: agents must be able to test, review, and validate their own work before deployment.
• Human engineers increasingly focus on planning, oversight, and system design rather than writing code line-by-line.
• Long-running optimization agents can continuously improve products, infrastructure, and development workflows.
#AIAgents #Cursor #SoftwareEngineering #AgenticAI #AICoding #DeveloperTools #LLMEvals #AIObservability #ArizeObserve
🔗 Try Arize AX & Phoenix OSS: arize.com
🔔 Subscribe for weekly content on LLMs, agents, and evaluation: youtube.com/@arizeai?sub_confirmation=1
What happens when one of the world’s leading AI coding companies turns its own product inward?
In this session, Cursor shares how its engineering teams use AI agents, automated evaluations, and agent-powered workflows to build Cursor itself. Learn how the company has evolved from AI-assisted coding to systems of specialized agents that plan, implement, review, test, and improve code with increasing autonomy.
The talk explores how Cursor structures its internal “agent factory,” including custom skills, automated code review agents, event-driven workflows, long-running optimization agents, and evaluation systems that enable scalable software development. You’ll also learn practical approaches to agent verification, CI/CD automation, risk-based code review, and continuous improvement loops that help agents learn from production feedback.
Key Takeaways
• AI coding is evolving from assistants to systems of specialized agents working together.
• Cursor uses custom skills, automations, and evaluation systems to improve agent performance over time.
• Verification is critical: agents must be able to test, review, and validate their own work before deployment.
• Human engineers increasingly focus on planning, oversight, and system design rather than writing code line-by-line.
• Long-running optimization agents can continuously improve products, infrastructure, and development workflows.
#AIAgents #Cursor #SoftwareEngineering #AgenticAI #AICoding #DeveloperTools #LLMEvals #AIObservability #ArizeObserve
🔗 Try Arize AX & Phoenix OSS: arize.com
🔔 Subscribe for weekly content on LLMs, agents, and evaluation: youtube.com/@arizeai?sub_confirmation=1
