AI Backend Engineer (f/m/d)

paretos is the leading AI-based Decision Intelligence platform for effective, data-driven decisions across entire enterprises. Our Business Agents make it possible to evaluate complex data volumes, predict scenarios, and derive optimal actions — from demand forecasting and replenishment to strategic C-level decisions. All through a code-free interface, with no prior data science knowledge required.
paretos is an inclusive company with equal opportunities for all. Our culture is guided by the principles of GUNG HO: meaningful work, shared goals, and mutual support. We are proud of a team of passionate, high-performing, and caring people of all genders, cultures, and backgrounds.

As AI Backend Engineer, you work in a hybrid setup, with 1–2 days per week in our Heidelberg office in close collaboration with Product, Forward Deployed Engineers, and Data Science colleagues. We look forward to your application if you already live in the Rhine-Neckar region or are willing to relocate.

Tasks

As an AI Backend Engineer, you will build the technical backbone of our Decision Intelligence platform. You develop scalable backend services that bring our forecasting and optimization models, our AI agent Socrates, and our semantic data layer into productive use. You work agentically and AI-first: Claude Code, MCPs, and custom skills are your standard toolkit. You think not only in terms of endpoints and databases, but in customer value, product impact, and adoption.

• You design, build, and operate scalable backend services in Python with FastAPI and Pydantic that productively support forecasting, replenishment, and agent workflows.
• You design clean API contracts between the platform, agents, ML models, frontend, and customer systems (ERP, BI, planning tools) and ensure robustness and observability.
• You build and extend our semantic data layer as the foundation for accurate forecasts and precise responses from our agent Socrates: modeling in PostgreSQL, transformations with dbt, pipelines, caching, and retrieval.
• You integrate LLM and agent systems deeply into the backend: tool use, function calling, RAG pipelines, eval frameworks for LLM outputs, guardrails, and production deployment.
• You work agentically and AI-first. Claude Code, vibe engineering, and subagents are part of your daily toolkit, and you continuously extend our internal setup (CLAUDE, hooks, skills, plan/execute/review loop).
• You translate business and customer requirements into viable technical solutions together with FDEs and Product, and make pragmatic trade-offs between speed, scalability, and maintainability.
• You think in terms of impact and have a strong product mindset: you measure whether your features actually achieve adoption, observe usage patterns, and sharpen the next iteration together with Product and FDE.
• You prototype quickly and close to the user: you put new ideas into the hands of FDEs and customers in days rather than weeks, gather early feedback, iterate closely on the actual workflow, and decide based on real usage what is worth hardening and rolling out.

Requirements

• 2 to 5 years of professional experience as a Backend, Software, or AI Engineer in a product-oriented environment — ideally in a product or SaaS context, startup experience welcome.
• Solid backend foundation: Python with FastAPI and Pydantic, clean API design, asynchronous programming, PostgreSQL and Redis, containerization with Docker, and experience with CI/CD pipelines.
• Strong product mindset: you don't just ask