Senior Deep Learning Scientist, Multimodal Agentic RL

Tasks

• Advance post training and alignment methods
• Benchmark model accuracy safety and task completion
• Collect and develop multimodal datasets
• Complete long horizon tasks in digital and physical environments
• Develop neural networks for language processing in agentic systems
• Enable planning and tool execution
• Implement RLHF and RLVR
• Optimize preferences
• Perform instruction tuning
• Research agentic reasoning and grounded perception
• Train fine tune and deploy advanced neural networks

Perks/Benefits

• N/A

Skills/Tech-stack

Dataset versioning | Deep learning | Distributed Training | Evaluation Pipelines | Experiment tracking | Human Feedback | Instruction Tuning | Learning from Human Feedback | MDP | Mixture of Experts | Multimodal Learning | Policy | Preference optimization | PyTorch | Python | Reinforcement Learning | Reinforcement Learning for Vision and Reasoning | Reinforcement Learning from Human Feedback | Reward Design | Transformer