RISE Lab
RISE Lab develops principled techniques for building reliable, secure, and efficient software systems. We study how program analysis can be extended with AI to reason about increasingly complex software, and how the same analytical foundations can be used to understand and optimize agentic systems. Our research spans program analysis, software engineering, and machine learning. We are based at the School of Computing, National University of Singapore, as part of PLSE@NUS.
Research Projects
Agentic Program Analysis
We develop program analyses augmented by large language models and AI agents, targeting problems such as bug detection, bug validation, and program verification. Models and agents help gather semantic context, formulate hypotheses, and guide analysis across multi-modal software artifacts. Our goal is to make program analysis more adaptive to new systems, properties, and software domains, while keeping its conclusions grounded in evidence that can be independently examined.
Program Analysis for Agentic Systems
Agentic systems span software tools, orchestration harnesses, and natural-language artifacts. Interactions among these layers introduce new challenges for observing, analyzing, and understanding the semantics of agentic systems. We develop program-analysis techniques tailored to different layers of the agentic stack, making cross-layer interactions observable and auditable. Our goal is to improve the reliability and performance of increasingly autonomous and large-scale agentic systems.
People
Research Highlights
Selected projects from RISE and our collaborators.
Join RISE
We are looking for Postdoctoral Fellows, Ph.D. students, research interns, and visiting students to join RISE Lab at NUS.
We welcome students with backgrounds in software engineering, programming languages, machine learning, and software security. Prior experience across all of these areas is not required. We are particularly interested in students who enjoy working on important problems in real software systems and developing principled technical solutions.