About the role
Spectral Labs is looking for a CAD / Geometry Engineer to build the geometry and data pipelines that power the next generation of CAD foundation models. Spectral released SGS-1, the first generative model for structured 3D parametric CAD, and is now building SGS-2. You will work at the intersection of computational geometry, CAD kernel internals, and AI research, collaborating directly with the research team on decisions spanning data representation and model architecture.
What will you be doing?
- Build high-performance geometry pipelines. Ingest CAD data across file formats, extracting and converting B-rep geometry, feature trees, and parametric data into universal feature representations for ML consumption.
- Work inside CAD kernels. Extract, manipulate, and reason about solid geometry programmatically.
- Shape the model with the research team. As the data representation evolves alongside the model, you weigh in on architecture and design decisions.
- Build automated extraction, encoding, reconstruction, and verification systems across the full pipeline.
- Lead data collection and curation, depending on experience.
About Spectral Labs
Spectral Labs is a spatial intelligence company building foundation models for the next generation of engineering design. Existing foundation models lack inherent spatial understanding and thus cannot meaningfully perform critical roles in engineering design and manufacturing. We train large models which lack these restrictions and can engineer physical systems. We believe our approach represents a significant step towards AI systems that can intelligently design and reason about physical systems.
