The "Customer Care" Department of Sber is looking for a Head of Data Science to lead the technical strategy for AI development.
The team is developing an AI solutions system for Sber's customer support, covering all customer interaction channels (> 21 million requests monthly), as well as key products for supporting the E2E support system: a unified knowledge base and customer journey design tools. The key mission is to build the most convenient and effective customer journey, relying on an advanced technology stack (including LLMs, agent systems).
Responsibilities:
- Formulating the strategy and ensuring results in the areas of Bank support development (AI assistants in channels, knowledge management, agent systems, LLM solutions quality assessment, etc.) in close collaboration with Business Leaders
- Defining architectural patterns for high-load AI services (chat, voice, asynchronous requests)
- Searching for and adapting the best global practices in NLP/LLM for the Bank's support processes and related products
- Improving current Agentic solutions, RAG solutions
- LLM alignment (Full Fine-Tuning, LoRA/QLoRA, RLHF/DPO/GRPO, etc.) for the specifics of banking support
- Optimizing LLM inference on CPU/GPU for high loads
- Mentoring TeamLead DS/DA specialists, conducting technical meetups
- Participating in hiring and assessing competencies in the D-vertical
- Further developing data-driven culture and infrastructure
- Interacting with product leaders at the Department Head level
Requirements:
- Deep theoretical knowledge and practical experience in NLP, LLM, DL
- 5+ years of experience in independently developing and fine-tuning models
- Excellent knowledge of design patterns for Agentic and LLM-included systems
- Deep knowledge of RAG mechanics, Alignment mechanics for LLM for business tasks
- Implementation of at least 2-3 end-to-end projects with core NLP/LLM technology in production, including the full integration cycle
- Understanding of LLM inference limitations and optimization methods in high-load systems, experience with monitoring and tracing tools for LLM applications
- Ability to generate and validate hypotheses, and present results in technical reports and presentations for C-level
- Practice in developing DS teams: conducting technical interviews, onboarding and developing Lead specialists, building individual development plans and architectural council systems
- Experience in managing technical teams
- Experience speaking at industry conferences or having publications in NLP/LLM will be a significant advantage
Will be a plus:
- Deep knowledge of GPU architectures, experience with distributed training
- Practice in building RLHF/DPO pipelines
- Experience with voice technologies (ASR/TTS) and multimodal models
- Experience in the banking sector and customer support domains
Conditions:
- Comfortable modern office near Kutuzovskaya metro station
- Work format - office
- Annual salary review, annual bonus
- Corporate gym and recreation areas
- Over 400 educational programs from SberUniversity for professional and career development
- Extended voluntary medical insurance, preferential insurance for family and corporate pension program
- Flexible discount on mortgage loans, equal to 1/3 of the Central Bank's key rate
- Free SberPrime+ subscription, discounts on products from partner companies
- Referral bonus for recommending friends to the Sber team
