Data Scientist (NLP|LLM)

SberHealth is a leading MedTech company in Russia, focused on improving healthcare through innovative technologies.

- Design and maintain the full cycle of improving medical LLMs: data collection, cleaning, versioning, training, and fine-tuning. - Build datasets and labeling contours: schemas and guidelines, consistency control, synthetic data generation, self-training, error and bias analysis. - Develop LLM-based pipelines and agents for medical tasks: RAG based on clinical recommendations, tool-calling, routing, multi-step workflows, orchestration (LangGraph and multi-agent frameworks), guardrails. - Create and develop an evaluation system: test sets and benchmarks, automatic metrics and LLM-as-a-judge where appropriate, expert validation with doctors, red-teaming, regression runs, A/B testing in production. - Conduct research iterations: formulate hypotheses, conduct experiments, perform ablation studies, document results, prepare scientific articles, and bring materials to publication.

- 3+ years in NLP/ML, proficient in Python: typing, testing, profiling, clean production code. - Practical experience in training and fine-tuning transformers: PyTorch + HuggingFace, understanding of Accelerate, DeepSpeed, or similar. - Experience in building data pipelines and reproducible experiments: datasets, versions, configs, tracking (MLflow or ClearML), ability to make correct comparisons. - Understanding of LLM systems: retrieval, tool-calling, agents, quality degradation, hallucinations, production limitations. - Quality assessment skills: metrics, benchmarks, error analysis, ablations, working with labeling and expert validation.

- Strong team of professionals passionate about their work; - Opportunity for growth in a leading MedTech company in Russia; - Cozy office in the City with a panoramic view of the city, hybrid work format; - Corporate equipment; - Medical program including telemedicine consultations, in-person visits to clinics, psychologists, dentistry, laboratory and instrumental diagnostics; - Paid English language courses; - Support for an active lifestyle — choose sports activities you enjoy (corporate squash, running, football in Moscow, and reimbursement of your sports subscription); - SberUniversity and payment for specialized training and courses.

СберЗдоровье is a Russian company providing digital medical services, including online booking to clinics and doctors, telemedicine consultations, diagnostic appointments, and corporate medical services. It operates as a leading aggregator with over 14 million monthly users, connecting patients to healthcare providers via the Sber app, and was formerly known as DocDoc before being acquired by Sber in 2017.