The role
- COMP
- $200K - $325K
- EQUITY
- Competitive
- LOCATION
- San Francisco
- WORKPLACE
- On-site
- EXPERIENCE
- 4+ years
- VISA
- None, Visa transfers
- STACK
- Python
- INDUSTRY
- AI, Data, Enterprise
/roles — ROLE_341
Machine Learning Engineer
Series B startup training vision models that make complex documents usable by large language models
The company
Agentic document platform used by leading AI teams to parse and process complex documents with enterprise performance at scale.
- STAGE
- scale-up
- FUNDING
- $100M+ raised
- TEAM
- ~80 people
- FOUNDED
- 2023
- BACKING
- backed by a16z, Y Combinator, Benchmark
JD — the work
About the role
The company builds vision models that interpret complicated documents much as a person would and converts them into structured input for LLMs. You would join a small machine learning group reporting to the CTO and co-founder, owning models from early research to deployment on actual enterprise workloads, from leading trading firms and major tech companies to AI teams that build on the platform. It suits someone who has personally taken a model from training to real users without handing it off.
What you'll do
- Train and ship leading models that extract structure and meaning from unstructured documents
- Test new methods for making LLM output more accurate, with tooling that proves whether each one helped
- Own the pipeline end to end: data preparation, evaluation and product integration
- Influence product and technical strategy by working with the founders and customers directly
- Talk technical details with enterprise customers as a regular part of the job
What they're looking for
- 4+ years of machine learning engineering experience
- Proven record of shipping models to production yourself, without handing them off half-done
- Strong Python skills and comfort working with unstructured, real-world data
- High agency, careful attention to detail and a refusal to settle for good enough
- Willing to work on-site full time in San Francisco