Jan 2020 – Jan 2022 · 2 yrs · San Francisco Bay Area

• Leading a centralized Data Science team driving impactful ML and AI solutions across the organization:
• Advanced Product Structuring: Automated tagging and attribute extraction for billions of products, surpassing industry benchmarks and significantly improving customer navigation.
• Demand Forecasting: ML-driven inventory predictions optimizing stock management.
• Revenue Optimization: Hybrid recommendation systems combining organic search with targeted ads, significantly boosting revenue.
• Scalable Content Moderation: AI-based detection of counterfeit, explicit, and brand-infringing content across massive datasets.
• Experimentation Platform: In-house A/B testing infrastructure enabling rapid, data-driven product iterations.
• Feedback Analytics: NLP and LLM-powered user feedback analysis directly shaping product strategy.

Jan 2017 – Jan 2020 · 3 yrs · Dubai, United Arab Emirates

• Leading a humble Data Science Team for Mobility of Things Domain.
• Making the most of out of AI applications for Multi-Stakeholder Marketplace. And yeah, the results are super yummy!
• A breakthrough ML dispatching system that has saved over 25%+ drivers' time globally
• AI personalization of the feeds: restaurants, cuisine, carousels and dish recommendations. The solution optimizes simultaneously for customers' conversion and experience, drivers time and marketplace fairness.
• Brand new ML Based tracking system, with 20%+ global improvement in the optimal ETA customer experience
• I joined the company as a senior data scientist, and at that time was developing solutions for the most exciting problems of a ride-hailing business:
• Demand-Supply matching, deep learning multi-objective optimization for subsidies allocation
• A spatial time series demand forecasting algorithm for surge pricing
• Deep learning demand and supply forecasting with an adjustment for extreme events (holidays, recurrent and asynchronous events) for long term planning
• Destination prediction algorithm which is used for drop-off location suggests and dispatching optimization
• Subsidy optimization (customers and drivers) to equalize demand and supply

Aug 2016 – Sep 2017 · 1 yr 1 mo · San Jose, USA

• Research and POC of a generative deep learning model (VAE, GAN) that learns hidden data center traffic structure and then generates similar traffic to use for simulation.
• Machine Learning techniques to analyze network traffic and develop a network modeling system. Classification, cluster analysis, anomaly detection, pattern recognition, time series analysis of network traffic.

Oct 2014 – May 2016 · 1 yr 7 mos · Moscow, Moscow City, Russia

• Designed and implemented realtime machine learning news popularity prediction system from the scratch. The system is based on Twitter data provided by Twitter Streaming API. The task was considered as a combination of regression and ranking problem.
• Worked with bank transactions to segment, cluster, classify bank's customers and identify anomalies in the transactions. Applied active learning approach for smart sampling.

Sep 2012 – Aug 2014 · 1 yr 11 mos · Moscow, Moscow City, Russia