Raghuveer Devulapalli
Senior Staff Software Engineer at NumPy
Hyderabad, Telangana, India
Summary
🤩
Rockstar🎓
Top SchoolRaghuveer Devulapalli is a Senior Staff Software Engineer and technical leader specializing in cloud software, high-performance x86 microarchitecture, and SIMD optimization, with seven years of focused industry experience and a PhD in Industrial Engineering. He is a key NumPy maintainer and the largest contributor of x86-based optimizations to NumPy, driving AVX-512/AVX2 math kernels and vectorized transcendental functions that delivered multi-fold speedups for widely used scientific workloads. Raghuveer founded x86-simd-sort—a C++ library now used in NumPy, PyTorch, OpenJDK, and Alibaba’s libGrape—and has applied similar performance engineering to AWS PostgreSQL and other cloud data services. His open-source impact spans onnxruntime, glibc, and Pixman, and his work earned Intel’s top honor, the Achievement Award. Based in Hillsboro, Oregon, he combines deep algorithmic and microarchitectural insight with hands-on engineering to boost throughput and accuracy in compute-intensive systems.
7 years of coding experience
8 years of employment as a software developer
Doctor of Philosophy (PhD), Industrial Engineering, Doctor of Philosophy (PhD), Industrial Engineering at University of MinnesotaMaster of Technology (MTech), Aerospace, Aeronautical and Astronautical Engineering, Master of Technology (MTech), Aerospace, Aeronautical and Astronautical Engineering at Indian Institute of Technology, Kharagpur
Stackoverflow
11reputation
1kreached
1answer
1question
Github Skills (17)
float3210
avx10
numpy10
performance-optimization10
simd10
floating-point10
numerical9
c-language9
algorithm9
algorithms9
numerical-methods9
cprogramming-language9
numeric9
vector6
sorting6
Programming languages (7)
Github contributions (5)
numpy/numpy
Mar 2019 - Sep 2022
The fundamental package for scientific computing with Python.
Role in this project:
Performance Engineer & Backend Developer
Performance Engineer & Backend DeveloperContributions:163 reviews, 143 commits, 91 PRs in 3 years 7 months
Contributions summary:Raghuveer focused on optimizing the performance of NumPy's float32 implementations for exponential, logarithmic, and trigonometric functions. Their work involved vectorizing these functions, particularly using AVX2 and AVX512 instruction sets and integrating with the SVML library, to improve accuracy and efficiency. They also fixed bugs in the AVX implementations, especially when dealing with special float values and improved testing for edge cases. Additionally, the user's contributions involved rewriting algorithms to leverage gather instructions and address compilation issues, improving the performance of NumPy operations.
pythonscientific-computingnumpy
numpy/x86-simd-sort
Oct 2022 - Oct 2022
C++ template library for high performance SIMD based sorting algorithms
Contributions:6 releases, 134 reviews, 2 commits in 1 day
cppsimdsorting-algorithmsargsortavx2
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.