Company: ghostbusters
Location: No limitations
Work Format: Remote
Salary: from 250 to 350 thousand rubles
The company is building an IBP / Integrated Business Planning platform for corporate planning in supply chains, logistics, and finance.
The team is developing an AI assistant for business planning that works with real corporate data and business logic: it analyzes internal data, runs forecasting and optimization algorithms, helps users with planning tasks, explains results, and identifies opportunities for improving business processes.
We are looking for an AI Engineer who will design and develop AI agents, agentic workflows, and LLM tools for planning / forecasting / reporting / automation.
Main Tasks
- Develop AI agents and multi-step agentic workflows.
- Integrate LLMs with APIs, databases, internal systems, and business logic.
- Work with RAG, embeddings, vector search, structured outputs, tool/function calling, memory/context.
- Configure evaluation: response quality, tool usage correctness, reliability, safety, performance.
- Improve prompts, tool definitions, and reasoning flows.
- Implement AI functions in production together with product / development teams.
- Communicate with business users and transform real planning processes into agent operations.
What's Important
- Strong Python and engineering practices: tests, CI/CD, Docker, Git.
- Experience deploying LLM / agent-based applications to production.
- Practical experience with agent frameworks: LangGraph, Claude Agent SDK, OpenAI API, or similar.
- Experience with APIs, backend services, SQL, and databases.
- Understanding of RAG, embeddings, vector search, prompt engineering, structured outputs, tool/function calling, and MCP.
- English B2+.
Will Be a Plus
- FastAPI / Django / PostgreSQL.
- Time-series forecasting: gradient boosting, time-series foundation models, and other approaches.
- Optimization engines: MILP, Pyomo, Gurobi, HiGHS, OR-Tools / CP-SAT.
- Experience with local / open-source LLMs.
- Experience using coding agents for rapid production-like prototypes.
Apply: recruiter contact
