Ivan Belonogov
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolIvan Belonogov is a deep learning and software engineer with eight years of experience applying ML to real-world products, from mobile face reenactment to production text classification. A competitive problem-solver—ACM ICPC 2017 World Champion and Google Hash Code 2018 winner—he combines algorithmic rigor with practical engineering. At VK he helped build YouTokenToMe, a high-performance BPE tokenizer that achieved significant speedups and practical adoption, and at Google he prototyped GAN-based semi-supervised fraud detection. Proficient in C++, Python, PyTorch and TensorFlow, he focuses on performance, quality (unit/integration testing) and deployment workflows including Docker. Based in London, he pairs a strong academic background from ITMO with hands-on system optimization and benchmarking expertise that often hides behind his contest accolades.
8 years of coding experience
2 years of employment as a software developer
Bachelor of Science - BS, Computer Science, Bachelor of Science - BS, Computer Science at ITMO University
Russian, English
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (13)
nlp10
c-language10
cprogramming-language10
python10
natural-language-processing10
testing10
dockers9
cli9
command-line9
benchmark9
benchmarking9
command-line-interface9
docker9
Programming languages (2)
Github contributions (5)
VKCOM/YouTokenToMe
Jul 2019 - Feb 2020
Unsupervised text tokenizer focused on computational efficiency
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:1 review, 21 commits, 15 PRs in 7 months
Contributions summary:Ivan primarily contributed to the performance testing framework and build system of the project. They added source code for speed tests, integrated docker for improved development workflow, and introduced support for selecting languages for benchmarking. Furthermore, the user added several unit and integration tests for the CLI and python API demonstrating a focus on quality assurance. They also refactored some of the underlying C++ code.
nlpword-segmentationtransformer-architectureunsupervisedtokenization
xbelonogov/YouTokenToMe
Jul 2019 - Feb 2020
Unsupervised text tokenizer focused on computational efficiency
Contributions:49 pushes, 23 branches in 6 months
nlptransformer-architectureunsupervisedfocusedtokenizer
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.