Saurabh Raje
PhD Intern at Apple
Salt Lake City, Utah, United States
Summary
👤
SeniorSaurabh Raje is a PhD researcher and software engineer with a decade of experience specializing in compilers, high-performance computing, and sparse linear algebra for large-scale scientific and ML workloads. He accelerates sparse tensor computations for quantum chemistry and has interned at Apple working on BNNSGraph compiler optimizations for Apple silicon, complementing prior research roles at IBM Research and ETH Zurich. His work spans applied research and production-focused engineering—optimizing SPMM kernels in nalgebra-sparse and extending DaCe’s TensorFlow frontend to generate high-performance, platform-aware code. Published at venues like ICML and Supercomputing, he combines deep systems knowledge with practical speedups (e.g., memory and training-time reductions) and a track record of award-winning research. Based in Salt Lake City, he brings both academic rigor and open-source impact to performance-critical ML and HPC toolchains.
10 years of coding experience
5 years of employment as a software developer
Github Skills (20)
algorithm10
algorithms10
matrix10
python10
algebra10
sparse-matrix10
data-structure10
vector10
tensorflow10
performance-optimization10
data-structures10
vector-math10
rust10
linear-algebra10
cluster-computing9
Programming languages (7)
Github contributions (5)
spcl/dace
May 2019 - Jul 2019
DaCe - Data Centric Parallel Programming
Role in this project:
Back-end Developer
Back-end DeveloperContributions:176 commits in 2 months
Contributions summary:Saurabh primarily contributed to the TensorFlow frontend, with a focus on expanding its capabilities within the DaCe framework. They worked on implementing support for various TensorFlow operations, including FusedBatchNorm, Conv2D, and other matrix operations. Their contributions include bug fixes and optimizations, as well as refactoring of existing code, such as the implementation of the Reshape transformation.
cudahigh-level-synthesisparallelhigh-performance-computingvivado-hls
dimforge/nalgebra
Feb 2022 - May 2022
Linear algebra library for Rust.
Role in this project:Back-end Developer
Back-end DeveloperContributions:2 reviews, 25 commits, 1 PR in 2 months
Contributions summary:Saurabh focused on optimizing the sparse matrix multiplication (SPMM) kernel within the nalgebra-sparse library. They introduced a new SPMM example and refactored the kernel, leading to significant performance improvements. Their contributions included pre-allocating memory, removing hash sets, and implementing other optimizations to speed up the matrix multiplication process. The user also addressed file handling for example.
eigenvalueslinear-algebra-librarylinear-algebrandarrayrust
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.