Intern for Sber's Data Warehouses and Analytics team. Research and development of AI agents for automating routine DWH tasks.
Tasks:
- Research and analysis of existing approaches to applying large language models and AI agents in the field of data warehouse management
- Formulating hypotheses, testing them, and developing prototypes of AI agents for automating routine data warehouse tasks
- Participation in creating pipelines for testing agents, primarily on synthetic data from data warehouses
- Evaluating the accuracy, speed, and security of agent operation
- Preparing data for model validation and training, reports, presentation of results and recommendations to the data warehouses and analytics team
Requirements:
- Knowledge of SQL (ability to write complex queries, understanding JOINs, window functions, optimization)
- Basic understanding of data warehouse architecture (star schema, ETL/ELT processes)
- Programming experience in Python
- Understanding of how large language models work: prompt engineering, calling models via API, fine-tune methods
- Ability to work with Git and Jupyter Notebooks / VS Code
- English language proficiency at the level of reading technical documentation (B2 and above)
Conditions:
- Official employment
- Internship for 3–6 months
- Flexible schedule (from 20 hours per week)
- Partial remote work
- Location: Moscow, Varshavskoe shosse (Chaika Plaza Business Center)
Skills:
- SQL
- Python
- Git
- Jupyter Notebooks
- VS Code
- LangChain
- LangGraph
- LangFuse
- GPU
- ROUGE
- BERTScore
