Company: Sber
Location: Moscow, RU
Employment Type: Full-time
ML Engineer in the AI Agents domain at Sber. Design and development of AI agents, deployment of ML models into production, working with data and pipelines.
Tasks:
- Design and develop AI agents: implement scenarios, orchestration, tool calling, and integrations with internal services
- Turn Data Science prototypes and models into reliable industrial solutions: package, embed into services and pipelines, version, and support
- Develop services, libraries, and pipelines in Python, write autotests, and participate in code reviews
- Design, build, and maintain data marts for ML models and agents
- Process large volumes of data using SQL, Hadoop, and Spark, optimize queries and computational tasks
- Ensure the quality and observability of solutions: logging, monitoring, error handling, and data quality control
Requirements:
- Industrial Python development experience of 2+ years
- Ability to write maintainable code: typing, testing, error handling, logging, and dependency management
- Proficient in SQL: complex queries, window functions, optimization, and working with large datasets
- Experience in building and maintaining ETL/ELT pipelines and data marts
- Experience in deploying services, models, or data products into production
- Understanding of AI agent principles: orchestration, tool calling, state management, and API integration
- Understanding of the ML model lifecycle — from artifact and data to inference, monitoring, and updating
Conditions:
- Work format: office or hybrid (Moscow)
- Full-time employment
Skills:
- Python
- SQL
- Hadoop
- Spark
- ETL
- ELT
- AI Agents
- tool calling
- LangGraph
- LangChain
- LlamaIndex
- LLM
- RAG
- vector databases
- Airflow
- Kafka
- Docker
- CI/CD
- Kubernetes
