Sergiy Matusevych
Principal Data Scientist
Redmond, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolSergiy Matusevych is a Principal Data Scientist with 14+ years building scalable, distributed ML systems and production-grade inference for audio and cloud services at Microsoft and prior research organizations. He blends deep research instincts with hands-on engineering—contributing low-level bindings for TorchSharp, data orchestration scripts for the DNS Challenge, and leadership of Apache REEF as a PMC chair. His work spans client-side audio models, petabyte-scale indexing, and auto-tuning frameworks (MLOS) for cloud performance, reflecting a rare mix of systems, ML, and functional-programming fluency. A polyglot developer and mentor, he moves seamlessly between C++, Java, Python, and Haskell and enjoys applying formal thinking to practical problems—from Project Euler puzzles to production DevOps. Known for thriving in hackathons and open-source communities, he brings both academic rigor (MS in Math/CS, studies at Stanford and Rutgers) and pragmatic delivery to complex, distributed ML challenges.
14 years of coding experience
29 years of employment as a software developer
Mathematics, Mathematics at Rutgers UniversityStatistics, Statistics at Stanford UniversityIgor Sikorsky Kyiv Polytechnic Institute
English, Ukrainian, Russian
Stackoverflow
473reputation
25kreached
6answers
4questions
Github Skills (24)
net10
bash10
dotnet10
blas10
asp-net10
libtorch10
dotnet-core10
compiler10
powershell10
csharp10
linear-algebra10
legend6
clojure6
cloud-services6
ggplot6
Programming languages (7)
Github contributions (5)
microsoft/DNS-Challenge
Oct 2021 - Feb 2022
This repo contains the scripts, models, and required files for the Deep Noise Suppression (DNS) Challenge.
Role in this project:
DevOps Engineer
DevOps EngineerContributions:2 reviews, 41 commits, 16 PRs in 3 months
Contributions summary:Sergiy primarily contributed to the repository by adding and modifying download scripts for various datasets related to the DNS challenge. They introduced scripts for DNS challenges 1, 2, and 4, including both bash and PowerShell versions, and made updates to existing scripts to reflect changes in data packaging and URLs. Their work involved managing data retrieval infrastructure and ensuring the availability of the necessary datasets for the challenge.
noisenoise-suppressionsuppressiondns
dotnet/TorchSharp
Nov 2018 - Nov 2018
A .NET library that provides access to the library that powers PyTorch.
Role in this project:Back-end Developer
Back-end DeveloperContributions:22 commits, 5 PRs, 2 pushes in 11 days
Contributions summary:Sergiy primarily contributed to the core library of `torchsharp`, focusing on low-level bindings for BLAS and vector functions. Their work included adding XML documentation to BLAS functions, updating structures, and implementing the necessary DLL import bindings. The user also fixed a bug in the `FloatTensor.Random()` function and performed some minor refactoring to improve code quality.
pytorchnet-librarydotnetpythonpowers
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.