Conduct applied research on transformer-based architectures for real-time recommendation systems. Publish findings internally, prototype models in PyTorch, and contribute to open-source initiatives. Budget for conferences and GPUs provided. 

We value TDD, code reviews, and continuous delivery.