Dmitry Razdoburdin
AI Frameworks Engineer at Intel Corporation
Munich, Bavaria, Germany
Summary
🤩
Rockstar🎓
Top SchoolDmitry Razdoburdin is an AI Frameworks Engineer with a PhD in Physics and Math and nine years of experience building high-performance algorithms and production software across optical communications, HPC, and ML. At Intel he tuned XGBoost and oneDAL for Intel CPUs and GPUs—delivering up to 10x training speedups on GPUs and sizable inference gains through SYCL kernel optimization and host-device synchronization reduction. His background includes developing ASIC-aware, fixed-point nonlinearity compensation and automated truncation optimization at Huawei, halving hardware complexity while preserving numerical accuracy. Comfortable in C++, SYCL and Python, he combines deep theoretical skills from academic hydrodynamics with pragmatic engineering to produce hardware-aware, deployable solutions. An active contributor to the flagship xgboost project, he has driven core histogram kernel refactors and SYCL inference integration that improved both performance and portability.
9 years of coding experience
4 years of employment as a software developer
Lomonosov Moscow State University
German, English, Russian
Github Skills (8)
gbm10
sycl10
xgboost10
machine-learning10
c-language10
cprogramming-language10
performance-optimization10
distributed-systems9
Programming languages (2)
Github contributions (5)
dmlc/xgboost
Sep 2022 - Oct 2022
Scalable, Portable and Distributed Gradient Boosting (GBDT, GBRT or GBM) Library, for Python, R, Java, Scala, C++ and more. Runs on single machine, Hadoop, Spark, Dask, Flink and DataFlow
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:62 reviews, 4 commits, 50 PRs in 1 month
Contributions summary:Dmitry's commits primarily involve optimizing the `BuildHistKernel` and related histogram building utilities within the `xgboost` library, focusing on performance enhancements and code refactoring. Their contributions also include implementing optional by-column histogram building and introducing a dispatcher for template parameters related to histogram building kernels. Furthermore, they introduced a new dispatcher for compile-time flags in histogram building, reflecting contributions to the core machine-learning logic. The user also appears to be involved in integrating SYCL support for inference.
xgboostpythonflinkdaskdataflow
razdoburdin/oneDAL
Apr 2023 - Sep 2024
oneAPI Data Analytics Library (oneDAL)
Contributions:1 PR, 37 pushes, 13 branches in 1 year 4 months
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.