Junior AI Agents Developer in Sber's Marketing Department. Under the guidance of a mentor, create agent systems on LangGraph for internal products and client conversational assistants based on LLM.

Tasks:

• Under mentor guidance, develop and maintain agent orchestration on LangGraph: implement graph nodes, work with state and conditional transitions
• Write and improve skills and tools for the agent: integration with external HTTP APIs, JSON Schema validation of results, simple ReAct subgraphs
• Work with LLM: write and edit prompts, design Pydantic schemas for structured output, experiment with model parameters and function/tool calling
• Develop HTTP interface: REST and SSE endpoints (dialog initiation, response streaming, history management) under the guidance of senior colleagues
• Participate in connecting external systems: HTTP APIs, Kafka, search engines; assist in developing the RAG pipeline (embeddings, search, reranking)
• Cover code with tests: unit and integration tests (pytest, LLM and external service mocks); adhere to quality standards — strict mypy typing, ruff linting, code review
• Participate in code reviews (initially as author, then as reviewer), learn the architecture of agent systems and banking security requirements
• Version and test prompts: compare behavioral scenarios, test on real models

Requirements:

• Proficient in Python (3.10+): OOP, data structures, basic algorithms; clean and readable code
• Practical experience with async/await / asyncio (commercial, internship, or serious pet project)
• Familiarity with FastAPI and/or aiohttp at the level of pet projects or commercial experience up to 1–2 years; understanding of REST API basics
• Understanding of LLM operation via API (GigaChat, ChatGPT, and similar): prompts, structured output, function/tool calling — at least at the level of personal experiments
• Familiarity with LangChain/LangGraph or other agent orchestrators, or willingness to quickly master them with a strong engineering foundation
• Knowledge of SQL and basic experience with PostgreSQL (queries, simple schemas, transactions)
• Use of Pydantic (or similar validation tools) and understanding the value of strict typing
• Willingness to write tests (pytest), use Git, follow code review and CI/CD standards
• High learning ability, interest in LLM technologies, and desire to grow in AI engineering
• Basic higher education is sufficient; we consider senior students and recent graduates

Conditions:

• Hybrid work format (2-3 days in the office), office at Kutuzovsky Prospekt, 32, Moscow
• Annual salary review, quarterly and annual bonuses
• Corporate gym and recreation areas
• Over 400 educational programs at SberUniversity for professional and career development
• Adaptation program and supervisor support at the start
• 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
• Prime subscription with the possibility of sharing with three close ones
• Reward for recommending friends to the Sber team