JOB-1628
A small, quickly growing startup is creating AI tools that speed up how physical products are designed in CAD. It seeks a flexible ML engineer who is comfortable switching between research ideas and production code.
Your focus
- Create, train and assess bespoke models and embeddings that power the design software
- Own the data pipeline, from collecting and curating data to the training framework
- Strengthen the machinery for training and testing, covering experiment logging, metrics and repeatability
- Carry models from prototype into live use, then monitor and iterate on real-world results
- Work with product and other teams to convert needs into ML features, and follow current research on LLMs and representation learning
About you
- Four or more years as an ML engineer in industry, or comparable academic research (research-based Master’s or PhD)
- Strong, clean Python with little need for hand-holding
- Track record inventing custom models and embeddings, rather than relying on off-the-shelf tools
- Hands-on LLM experience and fluency in PyTorch or a comparable deep learning library
- Experience taking ML to production quality
- Bonus: graph neural networks or transformers, cloud ML platforms, backend frameworks, or CAD and geometry exposure
On-site in the San Francisco Bay Area. H-1B transfers and TN visas can be supported; new H-1B sponsorship is not offered.
