Intern, Tool-Augmented LLMs for CAD/BREP 3D Object Generation

Description

• Prototype end-to-end workflows for turning natural-language descriptions into CAD/BREP objects using LLMs + external tools
• Explore retrieval-augmented generation (RAG) using prior CAD models, part libraries, or design templates
• Investigate how LLMs can interface with CAD geometry kernels (e.g., OpenCascade) to generate valid BREP and parametric 3D models
• Document research findings and develop actionable recommendations for integrating LLM-generated CAD designs into downstream engineering workflows
• Evaluate generative outputs across metrics such as accuracy, validity, manufacturability, and similarity to relevant design priors
• Develop tool-augmented reasoning pipelines that enable LLMs to call geometry APIs, validate shapes, and repair or refine generated models
• Implement iterative refinement and adaptive test-time compute methods to enforce geometric, functional, or manufacturing constraints

Requirements

• Knowledge of RAG, multi-step reasoning, or program-synthesis-style workflows is a strong asset
• Ability to work with structured geometric data formats (e.g., BREP, STEP, parametric constraints) is preferred
• Currently pursuing a Master’s or PhD in Computer Science, Artificial Intelligence, Machine Learning, Computational Geometry, or a related field
• Strong understanding of deep learning and/or LLMs, with experience in model prompting, fine-tuning, or tool augmentation
• Experience with Python, and familiarity with geometry processing, CAD APIs, or libraries such as OpenCascade, or similar
• Strong problem-solving skills and interest in applying generative AI to real CAD/engineering settings

Benefits

Training + Development

Interview process

Visa Sponsorship

Security clearance