Description
We are developing the MLOps platform and multi-agent SberWorks environment. Our area of responsibility includes managing the lifecycle of classical ML models, developing AI agents, and ensuring the operation of the AI function in OneWork. We are looking for a person to take on a research function: testing hypotheses, assessing the applicability of new technologies within our corporate architecture, and providing the team with ready-made solutions, not just theoretical findings.
The first stage of selection for this vacancy is a conversation with an AI recruiter. After applying, you will receive an invitation via email to take a preliminary interview with GigaRecruiter on Telegram. The dialogue will take approximately 10 minutes. Its goal is to clarify missing details and speed up the review of your candidacy.
GigaRecruiter is just starting its journey, so we ask for your understanding. Your experience and participation will help make it convenient and useful!
Responsibilities
- research new approaches and technologies in the field of classical ML and AI agents, assess their applicability within the existing tech stack and architecture
- quickly test hypotheses: from prototyping to deployed analysis with conclusions of "works / does not work / works, but with limitations"
- provide detailed feedback on experiment results: model accuracy, resource consumption, inference speed, compatibility with infrastructure, bottlenecks for implementation
- act as an internal expert and filter: weed out non-viable directions, highlight promising technologies before they go into full-scale development
- work closely with development teams and the department lead so that research results are immediately incorporated into the backlog and considered during planning
- participate in reviewing architectural solutions from the perspective of new components, assess the risks of introducing experimental approaches into the production environment.
Requirements
- 3+ years of Python development experience, confident command of the core ML stack (pandas, scikit-learn, PyTorch / TensorFlow, transformers)
- understanding of the ML model lifecycle: from experiments to deployment in production, experience with MLOps tools is a plus
- experience working with LLMs and agent approaches: understanding agent architecture, RAG, prompt orchestration, experience integrating large models into applied tasks
- ability to quickly understand new technologies and test hypotheses with code, not just "on paper." Prototyping and experimentation skills
- ability to work within existing architectural constraints: you understand that the corporate environment imposes its own requirements and know how to find solutions that fit within them
- strong analytical skills: willingness not just to say "this is cool," but to provide a breakdown by metrics, resources, and implementation complexity
- communication skills and teamwork: your results will directly impact development plans, it is important to be able to convey conclusions clearly and convincingly.
Conditions
- knowledge sharing with colleagues and a friendly team atmosphere
- work at Russia's largest bank on key projects
- employment according to the Labor Code of the Russian Federation
- regular corporate training
- voluntary health insurance (VHI), accident and serious illness insurance
- financial assistance and social support, corporate pension program
- preferential lending conditions.
