Maksim Volkov Email & Phone Number

Russia

Moscow, Moscow City, Russia

- Developing models that improved search page relevance and accuracy: 1) product text relevance determination, 2) product category filtering, 3) and filter ranking for search result pages.- Extracting and processing big data using PySpark from Hadoop for model training.- Conducted data exploration and visualization using pandas, matplotlib, and seaborn.- Collaborated with cross-functional teams to understand business requirements and deliver solutions.

Moscow, Russia

IDP team is a Competence Center in the field of Natural Language Processing and OCR at Sber. We provide NLP-related services at Sber and bank’s ecosystem and help other teams to improve their models and to translate the customer's requirements to the demands of the machine learning task.- Participated in development of AutoNER, a DL framework for document processing that accelerates NER and text classification model development by 50%. The framework is widely used within Sberbank and by more than 10 DS teams.- Led the development of a banking-specific entity extraction service used by multiple bank teams for document anonymization, customer document sorting, and other tasks. The use of the service reduces the time and resources needed for model development by DS teams.- Developed and presented models and services to business clients in processes such as SMS processing, loan document processing, and document management.- Curated a dataset for banking domain from public sources for pre-training and adapting pre-trained models such as BERT, LayoutLM v2/v3, and others. Also trained models for specific business tasks and hypothesis testing.- Built a document annotation pipeline, including instruction creation, communication with annotators, data validation, storage, and adapter writing for reading data from different sources.- Supervised interns.

Moscow, Russia

- Translated business requirements into machine learning models and presented research results to clients.- Conducted research projects on customer behavior, fraud detection, and insurance market analysis. 1) Proposed measures to save tens of millions of rubbles annually by analyzing the impact of updating customer contact information on future purchases. 2) Developed a fraud detection model to reduce the need for manual transaction monitoring, resulting in increased efficiency. 3)Made recommendations on customer loyalty and pricing policies based on research findings on mortgage customer behavior in the life insurance market.

Master'S Degree, Computer Science

Bachelor'S Degree, Economics

Master'S Degree, Mathematics And Statistics