Computer Vision Researcher

Stealth Deep Tech Startup | London Up to £150k + equity

About the Role

Most perception systems still rely on expensive hardware, specialist supply chains, and technology that has changed surprisingly little in decades.

This is a deep tech startup in stealth mode, backed by leading European investors, building novel perception systems for defence and security—pushing more capability into software and deploying it at the edge.

The team is currently seven people and growing, bringing experience across:

• Defence
• Deployed AI systems
• Security
• Hardware acceleration

They have real-world deployments at organisations such as Palantir and Helsing.

The work encompasses classical and modern computer vision, focusing on models running on severely constrained edge hardware, where power, latency, and memory are hard limits. Custom silicon acceleration is also part of their technical stack.

Responsibilities

• Design and develop models for image enhancement, processing, and understanding in degraded visual environments
• Optimise architectures for constrained edge hardware, balancing accuracy against latency, power, and memory
• Apply quantisation, pruning, and compression techniques to meet strict power and latency targets
• Build robust training pipelines and curate datasets capturing real-world edge cases and failure modes
• Work across a model family spanning:
  • Compact architectures
  • Transformer-based approaches
  • SSM (Mamba) models
• Collaborate with FPGA and hardware engineers to validate model–accelerator integration
• Translate field performance and failure modes into improved models and tighter training signals
• Help shape the research direction as the lab scales