Perception Engineer – Computer Vision & Edge AI | Zürich

About the Hiring Process

One application can unlock multiple opportunities. Our AI-powered matching platform considers your profile not only for this role, but also for other relevant opportunities across our client network.

We only advertise genuine, active opportunities. If you're a strong match, we'll be in touch. If not, we won't spam you.

About the Company

Our client is building the intelligent software layer powering the next generation of autonomous security systems. Combining AI, robotics, cloud infrastructure, and computer vision, they enable robots and intelligent sensors to detect, understand, and respond to complex real-world environments.

Backed by experienced founders, leading investors, and a world-class technical team, they're engineering the perception stack behind tomorrow's autonomous robotics.

• Location: Zürich, Switzerland (On-site | Remote considered for exceptional candidates)
• Employment Type: Full-time
• Level: Mid-Senior Level
• Visa Sponsorship & Relocation Support: Available

About the Role

As a Perception Engineer , you'll build and own the computer vision systems powering autonomous robots deployed in real-world environments. You'll transform state-of-the-art AI models into reliable production systems capable of detecting, tracking, and understanding objects across multiple cameras while running efficiently on edge hardware.

This is a builder's role where success is measured by real-world performance, not research papers. You'll work closely with AI, Robotics, and Software Engineering teams to deliver perception systems that operate reliably in demanding production environments.

What You'll Do

• Design and deploy cross-camera object re-identification (ReID) systems for people and vehicles

• Build high-performance object detection and multi-object tracking pipelines

• Integrate Vision-Language Models (VLMs) for scene understanding and operator intelligence

• Develop robust real-time video pipelines using RTSP and WebRTC

• Optimize AI models for NVIDIA Jetson and edge devices using TensorRT, quantization, pruning, and latency optimization

• Evaluate, fine-tune, and train computer vision models where required

• Build scalable image and video data pipelines for training and evaluation

• Collaborate closely with Robotics and Software Engineering teams to integrate perception into autonomous systems

• Continuously improve model reliability, performance, and deployment efficiency

What We're Looking For

• 3+ years of experience building and deploying production Computer Vision or Machine Learning systems

• Strong expertise in object detection, multi-object tracking, and cross-camera ReID

• Hands-on experience with Vision-Language Models (VLMs)

• Experience building real-time multi-camera video processing pipelines using RTSP and WebRTC

• Proven experience deploying AI models on edge hardware using TensorRT, quantization, pruning, or similar optimization techniques

• Strong Python and PyTorch (or JAX) skills

• Experience building reliable ML systems with strong software engineering practices

• Experience managing large image and video datasets for training and evaluation

• Builder mindset with strong ownership and the ability to thrive in a fast-moving startup environment

Nice to Have

• NVIDIA DeepStream

• MLOps (model versioning, CI/CD, monitoring)

• RAG and LLM-based systems

• Robotics, autonomous systems, surveillance, or defense

• Edge AI deployment

• Air-gapped or on-premise environments

• Publications in CVPR, ICCV, NeurIPS, or ICLR

What We Offer

• Competitive salary with meaningful equity

• Visa and relocation support

• Work alongside world-class founders and AI researchers

• Significant ownership over the perception stack from day one

• Opportunity to build production AI systems deployed in the physical world

• International, collaborative, engineering-first culture

• Direct impact on the future of autonomous robotics and intelligent security

Build the perception systems that enable autonomous robots to understand and navigate the real world...