Senior AI/ML Engineer

Senior AI/ML Engineer at SberSpasibo, Moscow. Develop AI agents for document search, improve RAG quality, and ensure security.

SberSpasibo is a bank loyalty program for Sberbank clients. The team develops AI agents for searching technical, regulatory, and accounting documentation. Tasks include improving RAG quality, evaluating agents, and configuring their protection. Ready-made models are used through a corporate gateway. Model training is not part of the tasks.

At least 2 years of commercial Python development At least one full-cycle RAG solution in production, experience with embeddings, chunking, re-ranking, and vector DBs: Qdrant, FAISS, or pgvector Understanding RAG architecture: metadata, embedding selection, source reference accuracy LLM system evaluation: reference sets, offline metrics, LLM-as-judge, regression checks, Ragas, DeepEval or similar Configuring protections against prompt injection and personal data leaks, validating response structure and content Prompt and context engineering in real projects, structured output and function calling Experience with at least one of LangChain, LangGraph, PydanticAI, OpenAI API or similar A/B tests of prompts and models in production, MCP or custom tool-layer for agents Basic SQL and relational databases

Salary not specified. Work in Moscow. [Details available at career.habr.com link]