Stanislav Kirillov
Lead Software Developer at Nebius
Berlin, Germany
Summary
🤩
Rockstar🎓
Top SchoolStanislav Kirillov is a Berlin-based lead software developer and founder with 11 years of experience building and scaling ML systems and infrastructure across Yandex, Toloka, and startups. He combines deep hands-on expertise in GPU-accelerated machine learning—demonstrated by CUDA bug fixes and refactors in the high-profile CatBoost project—with leadership roles overseeing ML systems and moderation platform development. At Yandex he progressed from developer to head of an ML systems group, driving performance and stability improvements for large-scale production models. Now leading teams at Nebius while founding InferScope, he blends engineering management with low-level performance work and a background in information security. Colleagues rely on him for pragmatic, efficiency-focused solutions that bridge research code and production-grade systems.
11 years of coding experience
8 years of employment as a software developer
Specialist, Information security, Specialist, Information security at Национальный Исследовательский Ядерный Университет «МИФИ» (бывший Московский Инженерно-Физический Институт)Лицей при КубГТУ
English, French
Github Skills (10)
cuda10
machine-learning10
c-language10
gpgpu10
cprogramming-language10
gpu10
accelerated-computing10
data-structure9
data-structures9
algorithms8
Programming languages (6)
Github contributions (5)
catboost/catboost
Oct 2017 - Jan 2023
A fast, scalable, high performance Gradient Boosting on Decision Trees library, used for ranking, classification, regression and other machine learning tasks for Python, R, Java, C++. Supports computation on CPU and GPU.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:21 releases, 23 reviews, 647 commits in 5 years 4 months
Contributions summary:Stanislav primarily focused on fixing and refactoring CUDA-related bugs and issues within the CatBoost library. Their contributions involved addressing bugs in CUDA code, including the model converter and testing files, ensuring correct bias handling and overall functionality. The user also worked on refactoring and simplifying code related to feature processing and data loading, specifically, with respect to sparse features and the use of iterators. This work was geared towards improving the efficiency and stability of the CatBoost library's GPU implementation.
kagglexgboostpythondata-mininglightgbm
kizill/ompi
Mar 2018 - Dec 2020
Open MPI main development repository
Contributions:22 pushes, 7 branches, 2 tags in 2 years 10 months
mpi
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.