Research Engineer - Vision & Vision-Language Models

As a core contributor on Zyphra’s Vision Team, this role focuses on building next-generation models for image understanding, computer use, and agentic visual tasks. The position involves working across the full model development stack, from data sourcing and generation to pretraining, post-training, reinforcement learning, and evaluation. The team owns its models end-to-end, offering opportunities to contribute wherever maximum impact can be achieved.