LLVM Discussion Forums

Hiring postdoc on new MLIR-related project

Hi everyone,

Prof. John Regehr and I (School of Computing, University of Utah) are looking to hire a postdoc to join our new NSF-funded project involving compilers, formal methods, and programming languages. More specifically, in the next several years we will be looking into formalizing the foundation of MLIR and building related tools based on that (interpreters, optimizers, translation validators).

Given the current situation, we are open to remote work.

Please contact me for more details or if you have any questions.

– Zvonimir

Thanks for the announcement: this sounds super exciting and highly needed!

I have very early work on formalizing some generic principles of the IR using axiomatic semantics, inspired by previous work and readings on memory models. For example, the recent discussion on relaxing dominance could be formalized by enforcing (or not) some relation within a customizable (pass-specific) happens before relation.

We have an intern, Magnus Lång who just started looking into this in a limitted setting.

It would be good to talk.