ML Engineer - 3D Geometry for CAD (m/w/d)

Kyrall · via Arbeitnow ·

TypeFull-time job
LocationMunich
Posted2 hours ago
We're building the world's fastest way to design real, functional 3D parts. Kyrall's AI-powered design agent turns ideas into ready-to-manufacture 3D models in seconds.
Backed by top European VCs, we're a young, well-funded startup based in Munich and we're just getting started.

The Role
As a Machine learning Engineer specialized in 3D geometry, you'll develop models that understand, manipulate, and generate CAD parts. You'll work on challenging problems involving 3D representations, geometric relationships and manufacturability.

What To Expect
Build and optimize ML models for 3D shape understanding and generation

Design data pipelines that process CAD formats into ML-compatible representations

Implement efficient training workflows for geometric deep learning

Deploy models to production and optimize for performance

Collaborate with product and engineering to integrate ML capabilities

Your Profile
2+ Years of working ML experience

Strong knowledge of PyTorch or TensorFlow

Experience with 3D data representation (B-Reps, point clouds, meshes, voxels, implicit surfaces)

Proficiency in Python and ML engineering best practices

Proven track record of bringing ML models to production

Nice To Have
Experience with CAD/3D software/libraries

Knowledge of computational geometry algorithms

Experience with MLOps tools and practices

Why Kyrall?
We're reinventing how machines are designed and built, from the ground up.
Our mission: build the world's first truly functional hardware design agent. 3D modeling tools were the last major leap in this space, but the way we design parts today has hit its limit. It's time for something radically better, and we're at the start of it.
As one of our first hires, you won't just execute the vision, you'll help shape it, with a competitive salary and stock options to match.
Find Jobs in Germany on Arbeitnow
ai
Apply on Arbeitnow →

Job sourced from Arbeitnow. Applications happen directly on the original platform — we never collect your data.