The role
- COMP
- $125K - $250K
- EQUITY
- Up to 1% equity
- LOCATION
- San Francisco
- WORKPLACE
- On-site
- EXPERIENCE
- 4+ years
- VISA
- None, Visa transfers
- STACK
- Python, PyTorch, TensorFlow, AWS, GCP, Azure, Django
- INDUSTRY
- AI, Hardware, Software Development
/roles — ROLE_342
Machine Learning Engineer
Small, fast-growing startup applying custom ML models to automate the tedious parts of hardware CAD
The company
Early-stage startup building AI-driven CAD tools that automate repetitive design work so hardware engineers can move from concept to final design faster.
- STAGE
- Series A-stage
- FUNDING
- $9M raised
- TEAM
- founding team of <10
- FOUNDED
- 2024
- BACKING
- VC-backed
JD — the work
About the role
The company builds CAD software with AI at its core to shorten hardware development, and its small team is growing quickly. This role owns the models at the heart of the product: designing custom models and embeddings, building the data pipeline, and taking work from early prototype into production. It suits a versatile engineer who likes wearing many hats and would rather invent new approaches than apply off-the-shelf tools. The team works on-site in San Francisco.
What you'll do
- Create, train, and benchmark bespoke models and embeddings at the core of the CAD product
- Own the data pipeline end to end, from wrangling and curating data to the training framework
- Improve the infrastructure for training and evals, from run tracking and metrics to reproducibility
- Move models from prototype into production, then monitor and iterate on real-world results
- Translate product requirements into ML features alongside design and engineering
- Write clean Python, review code, and bring in ideas from recent LLM and representation-learning research
What they're looking for
- 4+ years of industry ML engineering, or comparable research experience from a PhD or thesis master's
- Strong Python and the ability to work with little hand-holding
- Track record of creating novel models and embeddings yourself, rather than relying on prebuilt LLM or image tools
- Practical LLM experience plus fluency in JAX, PyTorch, or TensorFlow
- Comfort juggling several priorities in a fast startup, with strong problem-solving and communication
Nice to have
- Managed ML platform experience (Azure ML, SageMaker, or Google Cloud's AI tooling)
- Cloud platform experience on AWS or Google Cloud
- Backend development with Django, Flask, or SQL
- Background in CAD or computational geometry