Tiark Rompf

Tiark Rompf is an Associate Professor in the Department of Computer Science at Purdue University, where he has been a faculty member since Fall 2014. He is the founder and director of the Purdue Center for Programming Principles and Software Systems (PurPL), a research group focused on next-generation software systems that integrate data-driven algorithms, also known as "Software 2.0". His academic background includes a PhD from École Polytechnique Fédérale de Lausanne (EPFL), an MS from Universität zu Lübeck, and a BS from Universität Bremen, all in computer science.

His research spans programming languages, compilers, systems, databases, machine learning, and AI. He is a leading figure in meta-programming, domain-specific languages (DSLs), and compiler design, particularly through his work on Lightweight Modular Staging (LMS). His interdisciplinary work bridges theory and practice, aiming to make high-level programming languages performant and scalable.

The most recent publications highlight his focus on advanced type systems (e.g., reachability types), query languages (e.g., Flan, Rhyme), compiler IRs, and the integration of machine learning and data management systems. These works appear in top venues like POPL, OOPSLA, ICFP, SIGMOD, and ECOOP, indicating sustained influence and innovation in programming language research.

Among his scientific contributions, he has no explicitly listed awards in the provided text, but his leadership in founding PurPL and his extensive publication record in premier conferences are significant recognitions of his impact.

He advises a group of active researchers and students, many of whom are co-authors on his recent papers. His lab, PurPL, fosters collaborative research in programming principles and software systems, often involving meta-programming, differentiable programming, and system-level optimizations. He has led or co-led multiple collaborative projects, including work on Flare (a native compiler for Apache Spark), HACCLE (secure multi-party computation), and AutoGraph (imperative-style differentiable programming).