Role Overview
Join Siemens Digital Industries Operations Software to build AI-native industrial operations software. The role designs reusable industrial skills and maintainable workflows that connect AI, data, validation and production systems, with a focus on safe state changes and industrial correctness.
Responsibilities
- Design and implement versioned, reusable industrial skills and workflows for diagnostics, analysis and production actions.
- Define preconditions, parameters, result confirmation and failure behavior for skills that change system state; verify against test instances.
- Structure complex chains of skills, data access and validation into maintainable, testable workflows.
- Build interim MCP tools on the runtime where the skill engine lacks capability, then migrate when supported.
- Use coding agents AI-first while owning specification, review and industrial correctness.
Requirements
- Degree in engineering, computer science, automation or related field, or equivalent practical experience.
- At least three years developing software for manufacturing or process industries.
- Hands-on knowledge of a Siemens operations system such as Opcenter, WinCC, Teamcenter or Plant Simulation, or comparable MES, SCADA or PLM system.
- Solid Python skills; experience with workflow or low-code tools, APIs, data models and OPC UA information models.
- Practical experience with LLM-based applications, prompt design, tool calling via MCP or comparable approaches, agent frameworks such as LangGraph, vector databases or knowledge graphs, AI observability tools such as Langfuse, and Mendix or comparable low-code platforms.
- Ability to distinguish plausible from correct answers on the shop floor and test systematically.
- Fluent English; German is an advantage.
Nice to Have
- Experience with Siemens Workflow Canvas or comparable industrial workflow engines, quality engineering, root cause analysis, production planning, digital twins or simulation as validation tools.
- Experience with the NVIDIA NeMo stack or agent frameworks in production.