William Moses
Assistant Professor, Computer And Data Science
Cambridge, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolWilliam Moses is an assistant professor and researcher with 12 years of experience bridging compiler engineering, parallel computing, and machine learning systems. Based in Cambridge, he develops and optimizes compiler toolchains and parallel-code generation—contributions include LLVM/MLIR improvements and adding parallel CPU capabilities to Facebook Research’s TensorComprehensions. His work on Enzyme and LLVM shows deep expertise in automatic differentiation and thread-aware optimizations, and he has extended Enzyme to synthesize gradients for parallel and heterogeneous runtimes. He has held research roles at MIT CSAIL and DeepMind and applied high-performance computing expertise at places like SpaceX and Argonne National Lab. William combines rigorous academic training (MIT PhD) with production-grade systems work, notable for integrating low-level compiler intrinsics and runtime-level parallelism to speed ML workloads.
12 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy - PhD Electrical Engineering and Computer Science, Doctor of Philosophy - PhD Electrical Engineering and Computer Science at Massachusetts Institute of TechnologyHigh School Optics and Modern Physics, High School Optics and Modern Physics at Thomas Jefferson High School for Science and Technology
English, French
Github Skills (25)
c-language10
llvm10
compiler-development10
halide10
machine-learning10
intermediate10
automatic-differentiation10
multilevel10
parallel-computing10
compiler-optimization10
cprogramming-language10
julia10
optimization10
garbage-collection10
float329
Programming languages (24)
Github contributions (5)
EnzymeAD/Enzyme
May 2019 - Dec 2022
High-performance automatic differentiation of LLVM and MLIR.
Role in this project:
Back-end Developer & Compiler Engineer
Back-end Developer & Compiler EngineerContributions:159 releases, 1024 reviews, 1206 commits in 3 years 8 months
Contributions summary:William contributed to the Enzyme project, focusing on the automatic differentiation of LLVM and MLIR. Their work involved implementing features related to thread handling, including threadid and barrier instructions, and optimizing the LLVM-based code. Additionally, they worked on a custom compiler-like build system for the project, along with other improvements to the activity analysis and compilation process.
clangautomatic-differentiationsimulationcompilertensorflow
llvm/llvm-project
Feb 2021 - Mar 2025
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:Back-end Developer
Back-end DeveloperContributions:36 reviews, 56 PRs, 131 pushes in 4 years 1 month
Contributions summary:William contributed to the LLVM compiler infrastructure, specifically focusing on implementing and enhancing features within the MLIR (Multi-Level Intermediate Representation) framework. Their work included adding accessors for LLVM array types, enabling the import of inline assembly calls, and exporting parameter getters and setters for LLVM function types within the CAPI. Furthermore, the user addressed issues related to inlining and the folding of extract and constant operations to improve code optimization. The user's commits also show contributions to support alias scopes and floating-point value folding.
compilerstechnologiesclangsubmittoolchain
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.