Taxi Marketplace Efficiency | Dispatch & Driver Products TeamScooters | Analytics TeamResponsibilities:- Conducting and designing A/B experiments with network effects.- Managing the development team (backend, mobile, ML, QA)- Organizing product development through cross-functional teams consisting of analytics, ops, design, communications, etc.- Conducting research, ad-hoc calculations, and simulations before and after the launch of new projects- Mentoring new team members- Defining the product development strategyKey Projects:1. Led and managed the Taxi Airport Queue project, based on predicting the required number of drivers at the airport using ML and managing driver supply.Result: Reduced the number of waiting taxi drivers across the entire service geography at airports, and decreased the waiting time for a driver's order at the airport by 50%.2. The first analyst at Yandex Scooters, responsible for the efficiency of the scooter marketplace.Result: Developed initial reporting on key and real-time metrics. Implemented from scratch - pricing technologies, surge algorithms, scooter repositioning, and dispatching orders for battery replacements by couriers, and established safety riding zones.3. Was responsible for the analytics of marketplace efficiency for the second driver application, Uber Driver – a tool for attracting additional supply hours into the system. Result: Conducted experiments to modify the order distribution system, driver motivation, and penalty system, resulting in improvements in metrics such as acceptance rate, utilization rate, supply hours, and subsidy Gross Merchandise Value (GMV).4. Conducted and analyzed A/B tests with a network effect, evaluating various hypotheses for the order dispatch system in the main driver application.5. Implemented internal tools for analysts and managers.6. Combined two motivational driver tools, Activity and Priority, into a single one.Result: Raised order acceptance rate by 6%.

Yandex Maps• Used SQL-like query language to aggregate and analyze large amounts of information• Implemented MapReduce operations in Python to process data and measured different kinds of user metrics• Creating dashboards for product funnels in a mobile product• Clustered application users into different groups.

• Development of features for player profiles in PUBG, Dota 2 (Errors after team fights, weapon block, PUBG landing page prototype, etc.) • Implemented the Fantasy Prediction block (achieved the 100%-percentile in a real-time prediction competition, +12k sign-ups, and +100% DAU throughout the event, covered by the media).• Ad-hoc analytics for pro teams and media.