Currently I am a Compiler Engineer at Meta where I've worked on problems ranging from mobile-app optimization to automated SDK deployment. My work thus far has either shipped in production applications to millions of users or in one case made it into an LLVM lightning talk (and later used to deploy SDKs for another company's operating system, ie Fuchsia). I have spent a good portion of my time collaborating in efforts to do better engineering with upstream compiler codebases used at Meta. More recently I have been working on various efforts to help bring greater Swift adoption to Meta where I have largely been involved in working with framework teams to better interoperate their "ObjC++" heavy codebases with Swift.
I can be found at the LLVM Discord as puyanlotfi.
-
Apple: Worked on the team that engineered the Apple GPU compiler backend through many product ship cycles. I've worked on pre and post silicon bring up in various capacities. I built early iterations of internal Apple GPU tooling+infrastructure used for assembly, disassembly, profiling, and Machine IR (MIR) diffing; this partly involved the first effort to standardize assembly syntax for all tools for Apple GPUs across the entire company (assemblers, disassemblers, simulators, profilers, etc). Much of my work shipped with iOS, Xcode, went into upstream LLVM or an LLVM conference talk, was patented, or was used cross functionally across several orgs.
-
Intel: Briefly worked as a compiler engineer on "C for Metal". I am the original author of C for Metal's GPU printf, which is still the suggested debugging mechanism. I also worked on all the tools for verifying and disassembling the Intel GPU bytecode IR.
-
Georgia Institute of Technology: Got a BS and later an MS in CS.