Inside the Agent Engine: A Deep Dive with Traversal and LangChain
Traversal cofounders Anish Agarwal and Raj Agrawal join LangChain CEO Harrison Chase for a deep dive into the architecture behind AI systems for site reliability engineering. They discuss reasoning without labeled data, agentic search over petabyte-scale telemetry, online versus offline retrieval, context management, memory architectures, Production World Models™, agent evaluation, model selection, and designing reliable long-running AI agents.
00:00
/
00:00
How do you build AI agents that can reliably troubleshoot production systems at enterprise scale?
Recorded live at Traversal HQ in New York City, this discussion brought together AI builders to explore one of the hardest problems in enterprise AI: building agents that can reason reliably about complex production environments.
Harrison Chase (Co-Founder & CEO, LangChain), Anish Agarwal (Co-Founder & CEO, Traversal), and Raj Agrawal (Co-Founder, Traversal) discuss the architecture behind Traversal's AI SRE, from building a Production World Model and making enterprise telemetry searchable to designing agent harnesses, evaluating agents without labeled data, and continuously improving performance in production.