Traversal is an AI site reliability engineer for engineering teams responsible for large, interconnected production systems. It combines logs, metrics, traces, code, and deployment history into a live model of how those systems connect, then tests thousands of hypotheses in parallel to isolate an evidence-backed root cause and a remediation path in minutes. The company raised $120 million at a $1 billion valuation in a round co-led by Kleiner Perkins, Redpoint Ventures, and Khosla Ventures, with Sequoia Capital participating. The round comes as AI writes more of the code that reaches production, so changes ship faster and fewer people can explain a live failure. Customers running Traversal include Nubank, Harvey, PepsiCo, DigitalOcean, and American Express. It sits on the observability tools those teams already use, triaging alerts and tracing failures across services and infrastructure. Founded by causal machine learning researchers, including Columbia professor and CEO Anish Agarwal, the company is using the capital to move from naming a failure toward acting on it so software keeps running. The launch video shows that idea as a string of on-screen emergencies that end in seconds, because the investigation is already underway.