Working as an AI Engineer you will:

• Design and deploy LLM-based agents focused on planning and rollover tasks for a deterministic tax processing system.

• Connect intelligent agents to internal enterprise applications utilizing REST APIs and Model Context Protocol (MCP).

• Create scalable agent-to-agent (A2A) solutions capable of plugging directly into advanced, stateful orchestrators like LangGraph.

• Develop and optimize automated, Python-driven workflows within a live production environment.

Chabre IT Services is a global professional IT services provider, building long-lasting relationships with Enterprises. We specialize in the delivery of tailor-made solutions, smart outsourcing, try&hire, and success fee services. We are a smart IT boutique with unique knowledge, which will deliver your ideas into reality.

Our client is a globally recognized leader in professional services, financial consulting, and auditing. They focus on delivering innovative, technology-driven solutions to complex business challenges on a massive scale. By joining their team, you will contribute to an organization that heavily invests in secure, cutting-edge AI transformations tailored for the enterprise sector.

Must-haves:

• At least 1 year of commercial experience in creating Agentic workflows combined with solid enterprise-grade API integration skills.

• Advanced proficiency in Python, including asyncio, Pydantic for data validation, FastAPI, and testing tools like pytest.

• Proven track record of constructing LLM agents, focusing on structured outputs, multi-agent architectures, tool calling, and managing latency and costs (using frameworks like the Claude Agent SDK).

• Hands-on experience with Model Context Protocol (MCP) for setting up and consuming tool servers.

• Deep understanding of LangGraph for stateful orchestration, including handling human-in-the-loop workflows, durable checkpoints, and A2A communication protocols.

• Solid background in enterprise REST integrations, specifically utilizing secure authentication methods like OAuth2 and SSO (e.g., Azure AD).

• Familiarity with containerization (Docker) and fast in-memory data storage solutions (Redis / Redis Stack).

• Excellent Polish and English skills (at least C1 level).

Nice-to-haves:

• Knowledge of system observability and tracing tools (OpenTelemetry, Datadog, Langfuse).
• Experience with Azure cloud infrastructure, including blob storage and managed search.
• Familiarity with LLM evaluation techniques (e.g., LLM-as-a-judge) and spec-driven development.
• Prior exposure to the tax and accounting sector or existing familiarity with the client's internal technological ecosystem.

• Rate: 140 - 160 PLN/h + VAT
• Remote work
• Subsidy for peripherals in the amount of 500,00zł
• Working tool (MacBook Pro or Lenovo Legion 5)
• Co-financing of courses related to the position
• Benefits: MultiSport, Medicover