Ozon Bank

Ozon Bank is modern financial technology for people and businesses. We develop key systems ourselves: processing, accounting core, transfers. We offer the opportunity to influence the outcome, value openness, and support colleagues.

We are building an internal agent platform – a system that allows for the creation and management of AI agents for working with corporate knowledge. Agents connect to documents, tools, and services, helping teams get answers quickly and automate routine tasks.

The platform includes:

• Indexing and search of documents (Confluence, files, API sources, etc.) using vector databases;
• Management of agents and their tools (MCP servers);
• Integration with LLMs, embeddings, and internal company services;
• Administrative interface for managing entities and settings;

We are currently looking for a Project Manager who will oversee the platform's development, launch new features, and enhance existing tools related to agents, data, and AI.

Your responsibilities will include:

1. Leading platform development initiatives:

   • Gathering and prioritizing requirements from product teams, internal platform users, and information security;
   • Developing a roadmap for the quarter and year;
   • Decomposing and assigning tasks to the development team;
   • Working out architectural solutions together with technical leads;

2. Ensuring end-to-end delivery:

   • Synchronizing teams (backend developers, frontend developers, QA, ML engineers, infrastructure administrators);
   • Planning releases and monitoring the critical path;
   • Managing dependencies between services and external systems;
   • Performing risk management and ensuring status transparency for stakeholders;

Projects you will be working on:

• Agent Platform: Developing agent orchestration, working with tools (MCP), managing context and agent memory, request routing;
• Indexing and Search: Developing document indexing pipelines, chunking, generating embeddings, managing the Qdrant vector database; expanding the list of data sources and implementing hybrid search;
• Integrations and Infrastructure: Connecting new LLMs, MCP servers, internal company services, configuring observability (Grafana, Sentry), deploying to Kubernetes;
• Administrative Interface: Developing a UI for managing all platform objects;
• ML Component: Ensuring development and scaling to meet business needs, including working with models, integrating with vLLM and Triton for inference, and quality monitoring;

Our Tech Stack:

• Backend: Python, FastAPI, gRPC, SQLAlchemy, Alembic
• Frontend: JavaScript / TypeScript, Svelte
• Databases: PostgreSQL, Qdrant (vector database)
• AI/ML: Integration with LLM, embeddings, vLLM, Triton
• Infrastructure: Docker, Kubernetes (Helm), GitLab CI/CD
• Observability: Sentry, Grafana, Prometheus
• Task Trackers: Jira, Planner
• Documentation: Confluence

What's important to us:

• Experience in managing AI projects and full-cycle development projects (including backend, frontend, and QA) – from 3 years;
• Understanding of data and ML architecture: pipelines, vector databases, inference, monitoring;
• Ability to evaluate technical solutions and timelines, manage multiple initiatives in parallel;
• Skill in building and optimizing development processes with transparent reporting and prioritization;
• Strong communication skills: ability to synchronize teams, negotiate, manage risks and expectations;
• Experience working with internal customers and products;

Will be a plus:

• Understanding of the lifecycle of ML models and agent systems