Pinjia He

Associate Professor
The Chinese University of Hong Kong, Shenzhen国家级青年人才

Software Engineering × AI

We build capable, reliable software and AI systems—and help students grow into independent researchers.

Email Google Scholar GitHub DBLP X LinkedIn

About

I am an Associate Professor at the School of Data Science and the School of Artificial Intelligence, The Chinese University of Hong Kong, Shenzhen. I also serve as Assistant Dean (Research) at the School of Data Science.

My work has been published at top venues such as ICSE, FSE, ASE, ISSTA, ICML, NeurIPS, and ICLR, and has been cited 10,000+ times according to Google Scholar. I received the IEEE CS TCSE Rising Star Award in 2025 (the sole recipient worldwide that year). My work has received 7,000+ GitHub stars and 100,000+ downloads from 450+ organizations, and has been well recognized by industry, including Anthropic, Microsoft, Tencent, and ByteDance.

Before joining CUHK-Shenzhen, I was a postdoctoral scholar at ETH Zurich, mentored by Zhendong Su. I received my Ph.D. from The Chinese University of Hong Kong, supervised by Michael R. Lyu, and my B.E. from South China University of Technology.

My primary research interests lie at the intersection of software engineering and AI. I am also interested in CS and AI education.

Research Directions

Research overview

AI for Software Engineering

Building AI that helps people develop, test, and operate software.

Software Engineering for AI

Developing methods and tools to make AI systems reliable, secure, and effective.

Research Highlights

Publications

01 / ResearchAI for Software Engineering

OpenRCA

Evaluating how language models reason over system telemetry to locate the causes of software failures.

Diagnosing a failure requires connecting evidence across logs, metrics, traces, and system dependencies.

Anthropic's OpenRCA evaluation chart comparing Claude Opus 4.6, Opus 4.5, and Sonnet 4.5. — OpenRCA in Anthropic's Claude Opus 4.6 evaluation, February 2026.

ICLR25

Explore this work

02 / ResearchSoftware Engineering for AI

CipherChat

Evaluating whether language models remain safe when conversations use ciphers, revealing gaps in safety alignment beyond natural language.

Language models can understand encoded text even when their safety training is concentrated on ordinary language.

Diagram contrasting a model's refusal in a natural-language conversation with an unsafe response in a cipher-based conversation. — Communicating with GPT through ciphered text, then decoding its replies.

ICLR24 Stars630 585

Explore this work

03 / ResearchAI for Software Engineering

LogPAI

Open datasets and tools for AI-powered log analysis, from parsing raw logs to detecting system anomalies.

Understanding system behavior requires turning large volumes of unstructured logs into useful evidence.

IBM's log anomaly dashboard showing event counts and anomaly scores for a log template from one network device. — IBM adapted Drain for network outage monitoring.

Teaching

At CUHK-Shenzhen, I have taught Introduction to AI Programming (AIE1001), Introduction to Computer Science: Programming Methodology (CSC1001), and Software Engineering (CSC4001 and CSC6041).

Previously, I taught Research Topics in Software Engineering at ETH Zurich. During my Ph.D. at CUHK, I was a teaching assistant for Software Engineering and received the Excellent Teaching Assistant award.

Service

I serve as an Associate Editor of ACM TOSEM and Schedule Chair for FSE 2027, and served as Social Media Co-Chair for FSE 2025.

I co-organize EXPRESS 2026, the 2nd Workshop on Explainable and Reliable Software Systems.

My program and review committee service mainly includes ICSE, FSE, ASE, ISSTA, and OOPSLA.

Our Lab

Our mission is to advance software engineering and AI through rigorous, impactful research, while helping students become independent, capable, and dependable researchers.

Meet the Team Join Us