Arthur Suilin
Data Scientist
Belgrade, Central Serbia, Cyprus
Summary
🤩
Rockstar🎓
Top SchoolArthur Suilin is a data scientist and entrepreneur with over 20 years of experience building startups, analytics products, and high-performance data systems. He founded and exited two analytics-focused companies—RusNews (sold to Mail.Ru) and WebVisor (acquired by Yandex and integrated into Yandex.Metrica)—and later led Yandex.Metrica to become the world’s second-largest internet analytics service. A hands-on engineer, he contributed to ClickHouse and won a Kaggle web-traffic prediction competition, delivering the top solution and maintaining reproducible training pipelines. His technical breadth spans ClickHouse, MySQL, MongoDB, TensorFlow/PyTorch, Python/Java/Scala, and advanced methods like Bayesian analysis and Monte Carlo. Now based in Belgrade and working independently, he combines startup leadership, product-driven data science, and low-level database expertise to tackle large-scale analytics challenges.
14 years of coding experience
6 years of employment as a software developer
Bachelor of Science (BS), Biotechnology, Bachelor of Science (BS), Biotechnology at Российский Химико-Технологический Университет им. Д.И. Менделеева / Mendeleyev University of Chemical Technology of Russia
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (11)
kaggle10
time-series10
rnn-model10
tensorflow10
python10
n10
seq2seq9
modeling9
trainings9
machine-learning9
argparse9
Programming languages (4)
Github contributions (5)
Arturus/kaggle-web-traffic
Nov 2017 - Oct 2018
1st place solution
Role in this project:
ML Engineer
ML EngineerContributions:12 commits, 1 PR, 9 pushes in 11 months
Contributions summary:Arthur primarily focused on improving the training and submission code for a Kaggle competition on web traffic prediction. Their contributions included fixing boolean handling in argument parsing, correcting file names for data loading, and updating the code for compatibility with newer TensorFlow versions. These changes likely streamlined the model training and evaluation process, improving reproducibility and potentially the final results.
kagglernn-encoder-decoderrnnplaceseq2seq
Arturus/Metrika4j
Sep 2011 - Mar 2012
Contributions:11 commits in 6 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.