Description

We are an ML and AI team, part of the Sber development cluster. We are engaged in creating and implementing advanced agent and multi-agent systems, solving complex infrastructure tasks in the field of GenAI for corporate business. Our cluster consists of two teams focused on ML & AI tasks.

Our cluster develops AI Agents, ML models, and LLM applications for corporate business within the Corporate and Investment Unit, fully executing the entire work cycle: from data collection and analysis to development/implementation of models in PROD, monitoring, and defending solutions from Banking Regulators.

We work on high-complexity projects that require non-standard architectural solutions. We are looking for a Senior Developer who will take on the full cycle of creating agent systems.

Responsibilities

Main tasks:

• Development of agents: designing and implementing AI agents based on LLMs (typically, the SberDevice model line - GigaChat)

• Developing end-to-end integrations with external systems and other services via protocols such as Kafka, REST API, gRPC, WS, and others.

• Managing the agent lifecycle (agent lifecycle management): from setting the business task to release into production and subsequent support

• System design: designing the entire application architecture: from data sources and methods of their delivery to delivering results to the consumer, considering non-functional requirements (load, fault tolerance)

• Cross-functional interaction: working with IT teams and customers to align requirements, integrations, and remove blockers; essentially, you will act as the technical lead for your task, coordinating solutions with adjacent teams

• Solving infrastructure tasks: interacting with laboratory teams to organize and optimize the technology stack.

Requirements

Key requirements:

• Experience: solid experience in Python backend development (3+ years of commercial development), a portfolio with completed projects involving agents and/or agent systems

• Experience in roles such as ML Engineer, DS, MLOps, or Python Backend in ML

• Technologies (stack: A2A, MCP, LangGraph, LangChain, Kubernetes, Hadoop, Elastic, and many more.):

o Deep knowledge of frameworks: langchain, langgraph

o Understanding of RAG system architecture (retrieval-augmented generation) and experience working with RAG tooling

o Experience working with databases

o Understanding of containerization (Docker) and orchestration (Kubernetes) at a basic level

• Proficiency with tools for LLM monitoring and MLOps/ CI & CD

• Systemic thinking: ability to design an application architecture "from A to Z," anticipate bottlenecks, and make balanced technical decisions, ability to take a task from idea to production

• Personal qualities: proactiveness, excellent communication skills; you are not just "writing code," but can negotiate with adjacent teams, solve problems, and remove blockers; we value openness, willingness to share experience, and receptiveness to new knowledge.

Will be an advantage:

• Experience working in large corporations

• Experience in developing high-load agent systems.

Conditions

Conditions:

• Stable salary and social support for employees

• Official employment registration

• Extended DMS from day one and preferential insurance for family members

• Corporate training in Sber's Virtual School

• Free SberPrime+ subscription and discounts on products from partner companies

• Corporate pension program

• Flexible mortgage loan discount, equal to ⅓ of the Central Bank's key rate.