ML & AI Engineer with 8+ years of experience bridging advanced machine learning with real-world impact.

I'm an ML & AI engineer with over 8 years of experience bridging cutting-edge research with real-world business applications — from optimizing gas delivery routes and managing financial risks to developing graph neural networks and transformers for travel time estimation. With a Master’s degree in Mathematics from Moscow State University, I build production-ready solutions using Azure Databricks, Spark, and PyTorch, designing hybrid RL algorithms and API services to create scalable, long-lasting systems that deliver real value.

Experience
Data Scientist at Adaptfy (Jul 2022 – Present, Amsterdam, Netherlands)
Development and implementation of data & analytics solutions for the SHV energy family.

• Developed and adapted an efficient routing algorithm using the Open Source Routing Machine for gas delivery trucks in Belgium and the USA.

• Transformed existing solutions into more efficient, production-grade versions using Spark and Databricks.

• Worked extensively with Azure Databricks and Microsoft Azure.

Artificial Intelligence Engineer at Sberbank (2021 – 2022, Moscow, Russia)
AI research and application in Russia’s largest bank.

• Implemented graph neural networks for core banking activities.

• Conducted AI research in the graph neural network field.

• Translated research outcomes into stakeholder-ready analytics.

• Organized internal AI conferences and hackathons.

• Project highlight: “Logistics, Graphs, and Transformers: Towards Improving Travel Time Estimation.”

Data Scientist at Publicis Groupe (2019 – 2020, Moscow, Russia)
Machine learning for advertising and marketing.

• Led the development of marketing mix models to assess sales impact.

• Programmed ML pipelines in Python and R.

• Delivered modeling solutions for retail and tech clients.

• Built interactive dashboards with Flask and Shiny.

• Managed client-facing communications and model presentations.

Data Scientist at Russian Agricultural Bank (2015 – 2019, Moscow, Russia)
Risk modeling using classical ML methods.

• Built behavioral models to assess liquidity and interest rate risks.

• Automated monitoring of the bank’s securities portfolio.

• Created technical specifications for data flows and processes.

Projects
CayleyPy RL: Pathfinding & Reinforcement Learning on Cayley Graphs (Jan 2025 – Present)

• Built a hybrid RL and diffusion-distance algorithm for Cayley graphs with up to 10⁷⁰ nodes.

• Outperformed GAP benchmarks and proposed bounds related to OEIS sequence A186783.
  Technologies: PyTorch, Reinforcement Learning, Graph Theory, Discrete Math, Algorithms.

• Link - https://arxiv.org/abs/2502.18663v1

A Machine Learning Approach That Beats Large Rubik’s Cubes (Jun 2024 – Jan 2025)

• Created a neural network solution for diffusion-distance estimation and shortest-path search on 4x4x4 and 5x5x5 Rubik’s Cube graphs.

• Achieved over 98% optimality on 3x3x3 and 18x faster training compared to prior methods.
  Technologies: PyTorch, Graph Theory, Neural Networks, Algorithms.

• Link - https://arxiv.org/abs/2502.13266v1

TransTTE: Transformer-Based Travel Time Estimation (Jan 2021 – Jan 2022)

• Designed TransTTE, a graph-transformer model for spatio-temporal travel-time prediction.

• Built a working PyTorch prototype with enhanced accuracy.

• Presented at the PKDD 2022 Demonstration Track.
  Technologies: PyTorch, Graph Theory, FastAPI, Neural Networks, Cloud Computing.

• Link - https://arxiv.org/abs/2207.05835

Education
Moscow State University (2010 – 2015)
Master’s Degree in Mathematics
Location: Moscow, Russia