Description
We are the team of AI support assistants for Sber and its Ecosystem. Among our tasks is building and developing a system of end-to-end predictive events: predicting client inquiries, their goals, and preferred communication channels based on digital footprints (clickstream, interaction history, etc.). We handle 21 million inquiries monthly, creating the best customer experience.
We are looking for a Senior Data Scientist (with growth potential into a Lead position) who will join and lead the building and development of a scalable ML system.
Responsibilities
- Designing, developing, and enhancing an ML system: solutions for predicting inquiries (channel, need, time, intent, service scenario, etc.), including sequence-to-one / sequence-to-many approaches
- Working with event data: analyzing and modeling sequences of user actions (clickstream, inquiry and interaction history, logs)
- Researching SOTA approaches: conducting experiments with modern architectures for sequence modeling and comparing them with classic baseline solutions
- Production and scaling: participating in deploying models into production and integrating them into high-load decision-making systems (offline, near-real-time)
- Collaborating with business representatives and DS developers, mentoring colleagues
- Organizing validation and generating hypotheses to solve technical and business tasks.
Requirements
- 4+ years of experience, with a focus on RecSys, sequence classification, or behavioral analytics
- Sequence Modeling: deep understanding of working with temporal and event sequences, including their representation, aggregation, and model quality assessment
- Classic ML: confident proficiency in gradient boosting (LightGBM / CatBoost) and model calibration methods
- Excellent command of ML/DL, RecSys, and big data skills (Python, PyTorch, NLP, LightGBM / CatBoost, user modeling, sequence-based prediction, SQL, Hadoop, Spark)
- Willingness to take responsibility for architectural decisions, justify them to the business, and influence the product's business metrics.
Will be a plus:
- Experience building stable production pipelines, knowledge of Docker and model monitoring tools
- Experience designing A/B experiments, understanding mathematical statistics and methods for assessing result significance
- Experience with stream data processing (Kafka, Flink)
- Familiarity with NLP tasks
- Experience in fintech, banking, or support domains.
Conditions
- Comfortable modern office near Kutuzovskaya metro station
- Work format - hybrid possible after probation
- Annual salary review, annual bonus
- Corporate gym and recreation areas
- More than 400 educational programs from SberUniversity for professional and career development
- Extended voluntary health insurance, preferential insurance for family, and corporate pension program
- Flexible mortgage discount equal to 1/3 of the Central Bank's key rate
- Free SberPrime+ subscription, discounts on partner company products
- Referral bonus for recommending friends to the Sber team.
