معرفی
Pavel Panchekha is an Assistant Professor in the School of Computing at the University of Utah, where he holds the Warnock Chair for Junior Faculty. His research spans programming languages, web browsers, and numerical analysis, with a focus on developing programming language techniques to address challenges across computer science.
Dr. Panchekha received his educational training at prestigious institutions:
- PhD in Computer Science from the Paul G. Allen School for Computer Science and Engineering at the University of Washington, advised by Michael D. Ernst and Zachary Tatlock
- BS in Mathematics from MIT
Panchekha's research program has two major thrusts. First, he works on web browser internals, with projects including fuzzing layout invalidation, multi-tenant garbage collection, and optimizing 2D graphics. He is also authoring a textbook on web browsers that informs much of this research. Second, he focuses on automatic numerical analysis, with projects such as automatic accuracy improvement, synthesis via term rewriting, scalable static accuracy analysis, and math library implementation. He leads the FPBench and Herbie projects, which are major deployments of his research.
His scholarly output demonstrates consistent contributions across programming languages, verification, and numerical methods. Recent work shows a growing emphasis on bidirectional typing systems, layout invalidation in browsers, and robust floating-point error analysis. His publications reveal a trajectory from foundational work on floating-point accuracy (notably the Herbie tool that won a Distinguished Paper Award at PLDI 2015) toward more comprehensive systems for program synthesis, verification, and browser optimization.
Panchekha has received significant recognition for his research contributions:
- NSF Fellowship
- ARCS Foundation Fellowship
- Adobe Research Fellowship
- Wissner-Slivka Foundation Fellowship
- 2015 PLDI Distinguished Paper Award for work on the Herbie numerical analysis and repair tool
As an advisor, Panchekha mentors a substantial group of students across multiple levels. He currently advises six students: Marisa Kirisame (PhD), Bhargav Kulkarni (PhD), Yumeng He (PhD), Artem Yadrov (MS), Jesus Ponce (BS), and Jonas Regehr (BS). Previously, he has advised over twenty students including PhD candidates like Ian Briggs and numerous MS and BS students. His advising spans theoretical topics in programming languages and practical applications in web browsers and numerical computing.
Panchekha leads research groups focused on programming languages applications to web browsers and numerical analysis. His work on the Herbie tool for floating-point accuracy improvement has become influential in the programming languages community, and his more recent work on browser internals is shaping how researchers understand and optimize modern web rendering engines. He is currently developing a textbook on web browsers that aims to synthesize knowledge about browser architecture and implementation.
