Kiran Varaganti
Principal Member Of Technical Staff at AMD
Bengaluru, Karnataka, India
Summary
👤
Senior🎓
Top SchoolKiran Varaganti is a Principal Member of Technical Staff at AMD with 11+ years designing high-performance video codecs, computer vision systems, and linear algebra libraries. He architected AOCL-BLAS and contributed optimized microkernels to the BLIS ecosystem, bringing GPU and CPU performance engineering together. His background spans end-to-end video work—from MPEG2/H.264/SVC/HEVC codec development and frame-rate up-conversion to ML-driven video search, face recognition and gesture recognition. Kiran combines signal-processing rigor (M.Tech, IIT Guwahati) with practical multicore, OpenMP, CUDA and OpenCL implementations that move research algorithms into production. He has a track record of delivering complete decoders and reoptimizing core numerical routines, a blend that makes him equally at home in codec internals and sparse linear algebra. Based in Bengaluru, he is known for turning algorithmic complexity into highly optimized, production-quality software.
10 years of coding experience
15 years of employment as a software developer
M.Tech, Signal Processing, M.Tech, Signal Processing at Indian Institute of Technology, GuwahatiSt Joseph's Public SchoolB.E, Electronics and Communications, B.E, Electronics and Communications at Osmania University
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (12)
mat10
assembly10
hpc10
matrix10
blas10
x86-6410
assembler10
linear-algebra10
optimization10
x8610
c119
c179
Programming languages (4)
Github contributions (5)
flame/blis
Aug 2016 - Dec 2021
BLAS-like Library Instantiation Software Framework
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:31 commits, 2 PRs, 18 comments in 5 years 4 months
Contributions summary:Kiran implemented optimized microkernels for the BLIS library, specifically focusing on the trsm and axpyv routines. Their work involved modifying existing kernels, adding new optimized variants using techniques like FMA (fused multiply-add) instructions, and improving overall performance. These changes aimed to improve the efficiency of linear algebra operations, particularly for single and double-precision floating-point calculations. The user also reoptimized scalv routines.
software-frameworklinear-algebra-librarylinear-algebramatrix-multiplicationblis
kvaragan/clSPARSE
Aug 2015 - Dec 2015
Contributions:2 releases, 10 PRs, 78 pushes in 4 months
cudalibrary-softwaregpusparseopencl
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.