AI Agent Engineer & Data Architect (m/w/d)

AI Agent Engineer & Data Architect (m/f/d)

Full-time · Senior · Remote within the EU / Stuttgart · Start date by agreement

You build agent workflows and data integrations that run entirely within the customer environment, including orchestration, retrieval, evaluation, and system integrations.

BRANE increasingly supports automated workflows: agents that use tools, retrieve knowledge from customer documents, and complete multi-step tasks, all within the customer’s own environment. You will build these workflows and the underlying data architecture, including integrations with customer systems.

This role combines two disciplines: designing how enterprise data flows securely into AI systems and building the agent workflows that operate on that data. You will own both the concept and the implementation. One principle always applies: agents that cannot be measured do not go into production.

WHAT YOU WILL WORK ON

• Agent workflows, including planning, tool use, and self-correction

• Search and retrieval across customer knowledge bases

• Integration of customer systems and the underlying data architecture

• Evaluations that demonstrate the quality of these workflows

• Integration with the Guardian Core, ensuring that data flows and agents respect the policy and redaction layer rather than bypassing it

YOUR RESPONSIBILITIES

• Design and implement agent workflows using local models

• Build and maintain the integration layer for customer systems

• Make quality measurable before anything is deployed to a customer

• Improve retrieval performance using real enterprise data

YOUR PROFILE

• You have deployed agent-based or LLM workflows in production, not only as prototypes

• Hands-on experience with retrieval approaches and their evaluation

• Experience in data modelling and system integration

• Strong English skills for code and technical documentation

• Experience with vector databases and enterprise data access control models

• Experience with MCP or comparable tool-use standards is a plus

• Understanding of GDPR and the EU AI Act as design requirements is a plus

WHAT WE OFFER

• Two disciplines, one area of ownership: You design the architecture and build it yourself

• A demanding class of problems: secure and measurable agents operating on sensitive enterprise data

• What you build becomes part of a product that customers actively purchase, not an internal experiment

• A senior-level mandate: You make architectural decisions and take responsibility for them