Tech Stack: MySQL, Python

Expert’s cases:

1. Spearheaded the creation of a many-to-many matching algorithm for a novel bid/ask product, involving simulation-based evaluations of various approaches pre-launch. Contributed to the product's successful market entry and customer engagement

2. Engineered a pricing solution for bid/ask scenarios, optimizing price determination for customers and drivers, leading to reduced inefficiencies and enhanced customer satisfaction rates

3. Created a neural network based demand forecasting model (heatmap forecasting) based on recent studies in next frame prediction for video streams.

4. Led a unit of data scientists in Demand Side Platform (AdTech) (CPA-models, CTR-models, RecSys, A/B Testing)

5. Developed an antifraud system (ML-based antirobot model and business rules based on fraudulent domain detection), which increased our customers’ conversion rate by 3% on average by reducing the invalid or fraudulent traffic

6. Developed models for detection of resellers and promo abusers. Those solutions helped to save ca. 100K USD per month

7. Analysed fraudulent societies using Graph Neural Networks and User2Vec techniques, and increased the efficiency of fraud detection by 20%

8. Developed BI dashboards for management and data monitoring, which helped to detect data pipeline malfunction or data shift at early stages

9. Developed ML models and data pipelines for predicting customers’ needs (e.g grocery store visits) (Context Modelling), which increased overall conversion of push-communications (credit cards) by 2,6%