Senior Software Developer - Real-Time Platform & Edge Systems

PixelSight is a deep-tech startup that builds vision intelligence systems to tackle one of the hardest problems in perception: early detection of threats drones, boats, and other targets at the edge of what's visible. Our systems are deployed in the field and saving lives.

We're looking for a senior, hands-on software developer to own and evolve the real-time platform behind those systems the runtime that turns detection and tracking into one reliable system running on edge hardware, unattended, in the field.

Your ownership is the platform itself: process orchestration and inter-process messaging, latency and throughput budgets, camera and sensor I/O, deployment and updates on edge hardware, and the performance and reliability of the whole pipeline under real-world load. You'll integrate algorithmic modules into that runtime, and read and reason about their code well enough to debug the system end-to-end. The engineering satisfaction here is in taking something genuinely hard and making it run for months without anyone having to touch it.

Part of the role includes lower-level work on Linux-based systems close to the hardware, so systems experience is a strong plus. This is a good fit for someone who likes to take ownership, work across the stack, and help shape the product as it grows building meaningful, mission-driven technology that truly matters in the real world.

Requirements:

• 6+ years of software engineering experience, including 3+ years building and shipping production software- Must
• Production depth in a systems language: Python, C++ or Rust
• Independently design and deliver end-to-end features across a complex system
• Sound judgment on architecture trade-offs in real-time, resource-constrained environments
• Comfort navigating and improving a large, evolving codebase with minimal guidance

Bonus skills

• Multi-process architecture and IPC — message passing, shared memory, queues, backpressure
• Real-time performance work — end-to-end latency profiling, frame budgets, jitter, throughput under load
• Edge deployment — NVIDIA Jetson, JetPack, ARM64, cross-compilation, over-the-air updates
• Linux systems — systemd, networking, containerisation, debugging on constrained devices
• C++ and/or Rust — a significant advantage for real-time work
• Camera and sensor integration — industrial SDKs, ONVIF, RTSP
• Edge inference deployment — TensorRT / ONNX Runtime engine build, caching and versioning (deploying models, not training them)
• GPU performance — CUDA streams and memory management, profiling
• Domain — safety-critical, defence, aerospace, robotics or autonomous-vehicle software