Alexander Popkov
Sr. Data Scientist
Saint Petersburg, Saint Petersburg, Russia
Summary
🤩
Rockstar🎓
Top SchoolAlexander Popkov is a senior data scientist with eight years of hands-on ML experience spanning 2D computer vision, NLP and production ML system design, currently leading ML initiatives at Magnit Tech. He combines industrial engineering roles and team leadership with academic research—publishing in Q1/Q2 journals—and has a unique background applying image recognition to biological research at the Zoological Institute. Proven in production and open source, he contributed practical integrations to notable projects like SHAP (PyTorch examples and image plotting) and improved causal tree implementations in Uber’s causalml. Comfortable moving models from prototype to reliable systems, he also mentors small teams and bridges research-grade methods with business impact. Based in Saint Petersburg, he blends statistical training (Master’s in Business Information Analysis) with ecology-informed ML experience that informs robust, real-world modeling.
8 years of coding experience
8 years of employment as a software developer
Master's degree Business information analysis, Master's degree Business information analysis at Saint Petersburg State University
English, Russian, Japanese, German, French
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (22)
pytorch10
python10
scikit10
machine-learning10
causal-inference10
numpy10
lift10
deeplearning-ai10
explainable-artificial-intelligence10
deep-learning10
scikit-learn10
lifting10
modeling10
matplotlib9
statistical-models9
Programming languages (5)
Github contributions (5)
uber/causalml
Jun 2022 - Dec 2022
Uplift modeling and causal inference with machine learning algorithms
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 20 commits, 10 PRs in 5 months
Contributions summary:Alexander primarily contributed to the causal tree implementation within the `causalml` repository. They focused on updates and fixes related to the causal tree split criterions, bootstrapping, and `max_leaf_nodes` behavior. These changes involved modifying and optimizing the core tree structures for causal inference tasks, demonstrating a focus on enhancing the accuracy and functionality of the uplift modeling algorithms. The user also addressed potential issues and improved the code, including fixes and refactoring to ensure correct treatment effects and improve the overall quality.
causal-inferencemachine-learning-algorithmsmachine-learninguplift-modeling
shap/shap
Feb 2022 - Feb 2022
A game theoretic approach to explain the output of any machine learning model.
Role in this project:ML Engineer
ML EngineerContributions:11 reviews, 7 commits, 7 PRs in 2 days
Contributions summary:Alexander contributed to the `shap` repository by adding support for converting PyTorch tensors to NumPy arrays, demonstrating integration with deep learning frameworks. They also added functionality to save matplotlib images and implemented the option to include true labels in image plots. Furthermore, the user added an example using the Partition explainer with PyTorch, which showcases their knowledge of model explainability and integration of different tools.
machine-learning-modelsinterpretabilitymachine-learningdeep-learninggradient-boosting
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.