Towards an AIOps R&D Agent to Keep Critical Systems Working

Speaker: Professor Dan Pei
Department of Computer Science
Tsinghua University

Title: Towards an AIOps R&D Agent to Keep Critical Systems Working

Date: Friday, 20 November 2026

Time: 3:00pm - 4:00pm

Venue: Room 4472 (via lift 25/26), HKUST

Abstract:

This talk extends AIOps beyond IT operations to critical infrastructure, facilities, and equipment that must keep working as workloads and operating conditions change. We explore how locally deployable open-weight models can work with specialized algorithms, tools, and reusable skills to reason over metrics, logs, and traces. Representation, evidence, reasoning, and decision organize a design space for comparing method choices and reusing capabilities across tasks.

I will conclude with AIOps Code, a proposed R&D agent designed to run locally and automatically develop, test, and continuously improve AIOps agents to meet local operational requirements.

Biography:

Dr. Dan Pei is a Professor of Computer Science at Tsinghua University in Beijing, China. Previously, he was a Principal Researcher at AT&T Research. He received his Ph.D. in Computer Science from the University of California, Los Angeles, in 2005. His research focuses on AIOps (AI for System Operations), including failure detection, diagnosis, and reliable operation of complex systems. He founded the CCF AIOps Challenge in 2018, which has contributed to several widely used AIOps benchmarks.