Machine Learning Scientist I/II, Multi-Modal Scientific Reasonings
Core
Design and build state-of-the-art multi-modal reasoning systems using vision-language models (VLMs) to interpret scientific data including figures, plots, and microscopy images.
Role type
Research IC machine learning scientist (multi-modal scientific reasoning)
Builds
Production-ready systems for scientific superintelligence interpreting real-world scientific artifacts
Domain
Scientific research (materials, chemistry, physics, medicine) + Multi-modal AI
Deliverable
production ML models
Required skills
Multi-modal ML, VLMs, scientific QA/benchmarks, multi-modal fine-tuning, document parsing, dataset curation, benchmarking, PyTorch, Huggingface
Preferred skills
Microscopy image analysis, publications in top ML/CV/NLP venues, open-source contributions to multi-modal tooling
Technologies
PyTorch, Huggingface
Responsibilities
Lead research on multi-modal reasoning systems, design training/adaptation/test-time methods, build datasets and benchmarks, develop perception modules (OCR, table/structure recognition, plot parsing), collaborate with domain scientists and engineers
Seniority
Mid-to-Senior, hands-on IC
