Research Engineer
Hedra, San Francisco
About the job
ABOUT HEDRA
Hedra is a research and infrastructure lab advancing post-training and inference for multimodal and physical AI. Our ambition is to bring recursive self-improvement to models that learn from and interact with the real world.
We combine reinforcement learning research with GPU compute and optimized inference to help teams post-train, serve and continuously improve models they own.
We're backed by Index Ventures, a16z and Abstract Ventures. Our team works fully in person in San Francisco.
We're looking for people with startup experience who take ownership of their work and care deeply about the problems our customers need to solve. You'll help build the technology, products and business that turn this research into a lasting company.
WHAT YOU'LL DO:
- Design, implement, and run pre-training and post-training pipelines for action-conditioned world models and vision-language-action (VLA) models
- Develop and refine training methodologies, including fine-tuning, reinforcement learning, and large-scale multimodal learning
- Design and generate training and evaluation datasets from simulation, including environment setup, domain randomization, and sim-to-real transfer strategies