Reinforcement Learning for Vision Language Models
Core
Advance frontier vision-language models through reinforcement learning to improve visual understanding, reasoning, and multimodal capabilities.
Role type
Senior IC reinforcement learning engineer (vision-language models)
Builds
Vision-language models with enhanced visual perception, grounding, and reasoning capabilities
Domain
Artificial Intelligence, Machine Learning, Multimodal Systems
Deliverable
production ML models
Required skills
reinforcement learning, vision-language modeling, deep learning model training/evaluation, hypothesis formulation, experimental design, Python, PyTorch
Preferred skills
distributed RL training, synthetic data generation, reward model development, automated verifiers, visual grounding, document/chart understanding, image-to-code generation, spatial/physical reasoning
Technologies
PyTorch, Python
Responsibilities
Translate capability gaps into research hypotheses and measurable improvements; Develop targeted training tasks and synthetic data; Design and validate rewards, verifiers, and model-based judges; Build evaluations and conduct controlled experiments; Collaborate with scaling, data, and safety teams to integrate improvements
Seniority
Senior, hands-on IC