AI Engineer (m/f/d) for Chat-to-CAD / CAD-Automation

CADAICO; Stuttgart; Hybrid; EUR 60.000 - 85.000; Vollzeit

Benefits dieser Stelle

On-site in Stuttgart

Hybrid working

Flexible working hours

VSOP options

Zusammenfassung

CADAICO, a Stuttgart-based startup focused on AI-assisted CAD design, seeks engineers to build the AI core of a Chat-to-Design product. You will integrate LLMs and produce a production-ready CAD pipeline, learning from customer needs and enabling tailored design extraction from diverse inputs.

Join a small, high-caliber team and help craft reliable software features in a fast-growing startup environment with strong growth potential.

Qualifikationen

• Experience in using AI methods to solve problems in CAD design (2D/3D).
• Turning unstructured or ambiguous inputs into structured, reliable results – including non-textual sources.
• Excellent knowledge of Python and clean, maintainable engineering — coding, testing, APIs.
• Experience with retrieval (embeddings, vector search, re-ranking).
• Experience with Agentic Workflows: tool usage, function calling, multi-stage reasoning.
• A desire to delve deeply into a subject, make decisions and learn quickly.
• Very good English and good German language skills.
• Existing work permit for Germany.

Aufgaben

• Gradually develop the AI core of our Chat-to-Design product from LLM integration to a production-ready CAD pipeline.
• Ensure software learns from customers’ specific needs via prompts, knowledge bases, and memory; enable customization through architecture rather than weighting.
• Teach the product to reliably extract information from diverse inputs and feed into a CAD pipeline with retrieval and knowledge modules.
• Create Agentic Workflows: tool use, function calling, multi-stage reasoning and robust error handling.
• Implement complex AI ideas as reliable, productive software features in collaboration with engineers.

Kenntnisse

AI for CAD

Structured data extraction

Python

Embeddings / vector search

Agentic Workflows

English (strong)

German (good)

Work permit Germany

Tools

LangGraph

LlamaIndex

Pydantic AI

MCP

CAD software

Jobbeschreibung

Do you love using AI models to automate CAD design?

Welcome to CADAICO, the home for curious developers, AI nerds, geeks and engineering talents, with a focus on CAD design.

A large part of our work involves taking a wide variety of inputs – text, drawings, technical documentation – and extracting as much information as possible from them, feeding it into a CAD design pipeline. Tell us in two or three sentences where you’ve built something like this before: What were the challenges involved, and how did you solve them?

Our dev team is small but excellent – just two or three people. Your work is immediately visible and makes a difference from day one. If you’re looking for real responsibility, this is the place for you.

Our start-up

CADAICO, the ‘Chat-to-Design’ company, is a start‑up based in Stuttgart founded by Pedram, a former mid-level manager in mechanical engineering, and Daniel, an AI researcher from KIT.

Our team builds software that combines AI methods with conventional algorithms – enabling mechanical engineers to create their designs automatically, rather than manually using CAD tools.

With promising projects in the pipeline, we have a solid foundation for the next phase of growth and are looking for people who think independently, take responsibility and want to build real AI products that will shape the field of mechanical engineering.

What you’ll be doing with us

• Gradually develop the AI core of our Chat-to-Design product: from the LLM integration right through to the production-ready CAD pipeline
• Ensure that our software learns from customers’ specific needs – via versioned system prompts, knowledge bases, rule sets, feedback and memory systems. Achieve customer-specific customisation through architecture rather than weighting
• Teach our product to reliably extract all the necessary information from a small number of very different inputs – including non-textual ones – and feed it into a CAD pipeline, complete with the necessary retrieval and knowledge modules for design and component data
• Create Agentic Workflows: tool use, function calling, multi-stage reasoning and robust error handling
• Implement complex AI ideas as reliable, productive software features – in collaboration with our engineers.

Do you love using AI models to automate CAD design?

Welcome to CADAICO, the home for curious developers, AI nerds, geeks and engineering talents, with a focus on CAD design.

A large part of our work involves taking a wide variety of inputs – text, drawings, technical documentation – and extracting as much information as possible from them, feeding it into a CAD design pipeline. Tell us in two or three sentences where you’ve built something like this before: What were the challenges involved, and how did you solve them?

Our dev team is small but excellent – just two or three people. Your work is immediately visible and makes a difference from day one. If you’re looking for real responsibility, this is the place for you.

Our start-up

CADAICO, the ‘Chat-to-Design’ company, is a start‑up based in Stuttgart founded by Pedram, a former mid-level manager in mechanical engineering, and Daniel, an AI researcher from KIT.

Our team builds software that combines AI methods with conventional algorithms – enabling mechanical engineers to create their designs automatically, rather than manually using CAD tools.

With promising projects in the pipeline, we have a solid foundation for the next phase of growth and are looking for people who think independently, take responsibility and want to build real AI products that will shape the field of mechanical engineering.

What you’ll be doing with us

• Gradually develop the AI core of our Chat-to-Design product: from the LLM integration right through to the production-ready CAD pipeline
• Ensure that our software learns from customers’ specific needs – via versioned system prompts, knowledge bases, rule sets, feedback and memory systems. Achieve customer-specific customisation through architecture rather than weighting
• Teach our product to reliably extract all the necessary information from a small number of very different inputs – including non-textual ones – and feed it into a CAD pipeline, complete with the necessary retrieval and knowledge modules for design and component data
• Create Agentic Workflows: tool use, function calling, multi-stage reasoning and robust error handling
• Implement complex AI ideas as reliable, productive software features – in collaboration with our engineers.

Aufgaben

What you’ll be doing with us

• Gradually develop the AI core of our Chat-to-Design product: from the LLM integration right through to the production-ready CAD pipeline
• Ensure that our software learns from customers’ specific needs – via versioned system prompts, knowledge bases, rule sets, feedback and memory systems. Achieve customer-specific customisation through architecture rather than weighting
• Teach our product to reliably extract all the necessary information from a small number of very different inputs – including non-textual ones – and feed it into a CAD pipeline, complete with the necessary retrieval and knowledge modules for design and component data
• Create Agentic Workflows: tool use, function calling, multi-stage reasoning and robust error handling
• Implement complex AI ideas as reliable, productive software features – in collaboration with our engineers.

Qualifikation

What you should bring to the role

• Experience in using AI methods to solve problems in CAD design (2D/3D)
• Experience in turning unstructured or ambiguous inputs into structured, reliable results – ideally also from non-textual sources (images, documents, technical documentation)
• Excellent knowledge of Python and clean, maintainable engineering – coding, testing, comprehensible APIs
• Experience with retrieval (embeddings, vector search, re‑ranking) – and an understanding of why this is more challenging with real-world data than in tutorials
• Experience with Agentic Workflows: tool usage and function calling, multi-stage reasoning – and an understanding of what happens when the model gets it wrong
• A desire to delve deeply into a subject, make your own decisions and quickly get to grips with unfamiliar topics
• Very good English and good German language skills for confident internal and external communication
• An existing work permit for Germany. As a small team, we’re unfortunately unable to provide visa support.

It’d be great if you know a bit about these – you’ll learn the rest with us.

• Agent frameworks such as LangGraph, LlamaIndex, Pydantic AI or MCP
• Systematic evaluation of LLM outputs (Golden Sets, LLM-as-Judge, Langfuse, Ragas)
• Experience with extracting information from images or documents
• CAD software or experience in the mechanical engineering field

Benefits

What we offer

• Based in Stuttgart (on-site) with highly flexible working hours and hybrid working
• Agile working with flat hierarchies: No endless approval loops. You’ll take responsibility, make decisions and implement them quickly
• Strong development opportunities: Your opportunities grow alongside our start-up
• Genuine responsibility and a visible impact: What you do matters. A great deal, in fact.
• Room to experiment: Develop ideas, test them quickly, learn and put them into practice
• Salary: €60,000 – €85,000 + VSOP options depending on experience and level of responsibility.

Why CADAICO?

If you want to take on responsibility and enjoy going to work, you’ve come to the right place ;)

We welcome applications from people of all backgrounds, age groups and identities.

We look forward to getting to know you.

Pedram