We're hiring a few people, and are looking for research-minded engineers to build real-world AI for safety- and privacy-critical settings, specifically policing. For police to serve people in a transparent and consensus-based way is demanding for many reasons. Lack of good interactive tools should not be among their challenges.
The roles span from cloud training infrastructure to inference on low-power edge devices. We want people who think like scientists: you form hypotheses, design experiments, build or stress-test benchmarks, and make progress while mitigating for uncertainty.
Several roles are open, and for strong candidates with range, the lines between them are negotiable:
* Vision-language models - data preparation, architecting, training, and applied research.
* Vision model distillation & edge deployment - training and shipping models for low-power devices.
* Cloud engineering for ML - training and inference infrastructure at scale.
A computer-vision background is one natural fit. But we're working on multimodal models, so your ability to absorb and apply new findings might show up on your CV in other ways.
Published, shipped, or both? If you like hard engineering problems end-to-end, let's talk.
(Team is partly in central London and partly in Amsterdam, so it's ideal if you're local or open to relocating)
If you're excited by this but not sure you fit, please reach out anyway. We'd rather hear from you than have you count yourself out.