I’m a PhD Candidate at UC Irvine, working on problems related to software dependability (specifically, security, reliability, and safety) in smart systems. I propose to ensure software dependability in smart systems by (1) conducting comprehensive studies of bugs and vulnerabilities found in such systems, (2) automatically generating effective and valid tests, and (3) leveraging static and dynamic analysis techniques to detect, reproduce, and repair bugs found in smart systems.
To this end, I conducted a longitudinal study on over 7000 apps from GooglePlay to identify vulnerabilities in third-party native libraries in Android apps. The goal of this work is to study: (i) the prevalence of vulnerabilities in native libraries, and (ii) the speed at which app developers apply security patches. This work has been accepted into ICSE 2021. My current work continues along this path involving smart systems, where I created a software test generation approach for autonomous vehicles (i.e., self-driving cars) that uses evolutionary algorithms with a novel gene representation and test oracles to determine safety and motion sickness-inducing violations. I received the SIGSOFT Frank Anger Award 2021 for a proposal version of this work; a full version of this work is under review. This testing framework builds upon a comprehensive study we conducted of bugs in autonomous vehicle software systems—categorizing them in terms of bug symptoms, causes, and the component types they affect. This work is published at ICSE 2020.
I’m a research intern at RiSE in Microsoft Research. I work with Madan Musuvathi and Shan Lu on performing empirical analysis of high impact Azure cloud incidents to understand the root causes of bugs behind such incidents and their fix patterns. The goal of this work is to provide data-driven recommendations to other teams and partners to improve the effectiveness of current tools, or suggest new tools and extensions for input validation, fault handling, API usages, handling leaks, etc.