Dhiraj Kalamkar
Principal Engineer at Intel Corporation
Bengaluru, Karnataka, India
Summary
👤
Senior🎓
Top SchoolDhiraj Kalamkar is a Principal Engineer at Intel’s Parallel Computing Lab in Bangalore with 11+ years driving R&D on parallel and GPGPU architectures to accelerate deep learning on Intel Xeon and GPU platforms. He brings rare full-stack expertise across AI workloads (Vision, DLRM, LLMs), frameworks and runtimes, low-level libraries, and hardware, and has shaped low-precision DL roadmaps and security features for Xeon processors. Dhiraj has a strong track record in benchmarking and performance engineering, contributing to multiple MLPerf submissions that set leading training and inference results on Xeon. He is an active contributor to open-source projects such as libxsmm, where he improved TensorFlow LSTM support and delivery tooling, and holds six patents alongside 15+ peer-reviewed publications. Based in Bengaluru, he blends academic rigor from IIT Kanpur with hands-on systems optimization to turn research into production-grade performance. An often overlooked strength is his cross-layer debugging ability—tracing performance issues from model down to microarchitecture.
11 years of coding experience
17 years of employment as a software developer
Primary, Primary at Balak Mandir, JinturHSC (12th), Science, HSC (12th), Science at Mahatma Gandhi Mahavidyalaya, AhmedpurIndian Institute of Technology KanpurBE, Computer Science and Engineering, BE, Computer Science and Engineering at Government Engineering College, AurangabadSSC, High School, SSC, High School at Jawahar Vidyalaya, Jintur
Github Skills (9)
matrix-multiplication10
machine-learning10
lstm10
tensorflow10
sparse9
python8
avx7
jit6
fortran5
Programming languages (3)
Github contributions (5)
libxsmm/libxsmm
May 2019 - Sep 2022
Library for specialized dense and sparse matrix operations, and deep learning primitives.
Role in this project:
ML Engineer
ML EngineerContributions:7 reviews, 8 commits, 4 PRs in 3 years 5 months
Contributions summary:Dhiraj contributed to the TensorFlow wrapper code for LSTM operations, specifically focusing on the backward pass (gradient calculation) within the `tf_lstm_ops` directory. They added a setup file for creating a wheel package, indicating involvement in building and distributing the library. The user also addressed merge conflicts and added a sparse adagrad kernel reproducer and a run script for the same.
bfloat16avxsimdlapackavx512
ddkalamk/libxsmm
May 2019 - Jul 2024
Library targeting Intel Architecture for specialized dense and sparse matrix operations, and deep learning primitives.
Contributions:12 pushes, 2 branches, 1 tag in 5 years 3 months
cudasparse-matrixdeep-learningsparseintel
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.