Senior Data Scientist
Current- Created a streamlined repository of recommender models, facilitating the efficient development of new pipelines with reduced code complexity, improving team productivity.- Introduced the practice of offline measurement for recommendation models prior to A/B testing, enabling early rejection of suboptimal models.- Organized and led a bi-weekly reading group for Data Scientists, attracting 20-30 participants to discuss advancements in machine learning and data science.- Designed a neural network model (M2) based on the encoder-decoder architecture, utilizing Samokat data, inspired by relevant research.