The role
- COMP
- $150K - $220K
- EQUITY
- Competitive equity
- LOCATION
- San Francisco
- WORKPLACE
- On-site
- EXPERIENCE
- 0 - 2 years
- VISA
- None, Visa transfers, New visa sponsorships
- STACK
- Python, C/C++, PyTorch, Git
- INDUSTRY
- AI, Manufacturing, Robotics
/roles — ROLE_452
Engineer, New Grad
Research-heavy startup whose AI models learn to engineer physical parts directly from CAD data
The company
AI lab training large foundation models with a native grasp of 3D geometry, aimed at automating engineering design and CAD work for physical products.
- STAGE
- early-stage
- FOUNDED
- 2022
JD — the work
About the role
This is a deliberately broad early-career role at a lab whose foundation models are meant to design physical systems. The team hires new graduates for raw talent over domain background, and the work mixes AI, ML, graphics and geometry, from data pipelines and eval harnesses to internal tools and support for training runs. You should be genuinely strong in at least one of these fields and curious about the rest, and your scope expands as you earn trust.
What you'll do
- Develop pipelines that read CAD files, translate between formats, pull out features, and confirm nothing was lost
- Create eval harnesses and tools that reveal whether a new model is truly an improvement
- Support experiments and training jobs by adding instrumentation, tracking down failures, and summarizing results clearly
- Create internal tooling that makes daily work smoother for your colleagues
- Own clearly defined slices of bigger research projects, taking on more responsibility as you prove yourself
What they're looking for
- A recent graduate with 0 to 2 years of experience
- Real depth in ML, geometry, graphics or AI, plus curiosity about the other areas
- Comfort with the team's core tools: Python, C/C++, PyTorch and Git
- Strong raw ability and fast learning, since domain experience is not expected