(Senior) Machine Vision Engineer

You will join the Inventory Insights squad at Verity, working on the perception stack that turns raw 2D and 3D drone sensor data into reliable inventory signals such as barcode reads, occupancy estimates, pallet segmentation, and item counts. You will adapt, fine-tune, and evaluate open weight models for warehouse environments, build robust solutions for real-world conditions, analyze failures on real drone data, and optimize models for deployment on drone hardware or cloud inference. You will also spend time on-site with clients to understand warehouse operations and translate ambiguous operational needs into tractable perception problems.

Responsibilities

• You'll focus on the perception stack that turns raw 2D and 3D drone sensor data into reliable inventory signals: barcode reads, occupancy estimates, pallet segmentation, item counts.
• Adapt, fine-tune, and evaluate open weight models (detection, segmentation, VLMs) for warehouse environments
• Build robust solutions for real-world conditions: occlusion, bad lighting, reflective packaging, partial views
• Analyse failures on real drone data and close the loop back to production
• Optimize models for deployment on drone hardware, cloud inference, or both
• Spend time on-site with clients to understand how warehouses actually run, then turn ambiguous operational needs into perception problems that are both technically tractable and genuinely valuable to solve

Requirements

• MSc or PhD in CS, robotics, ML, or computer vision (or equivalent experience)
• Several years of hands-on experience building CV or ML systems that shipped to real users
• Strong practical depth in deep learning for visual perception: training, fine-tuning, evaluation, failure analysis
• Experience in at least some of: object detection, segmentation, depth estimation, OCR, 3D perception, or multimodal vision models
• A product mindset: you get close to client needs, scope problems for feasibility and impact together, and care that what you ship measurably improves operations rather than just benchmark numbers
• Solid Python and comfortable in PyTorch or TensorFlow
• Clear communicator in English

Nice to have

• Open weight foundation models for vision or multimodal reasoning
• C++
• Edge deployment: ONNX, TensorRT, quantization, edge GPUs
• RGBD data, time-of-flight cameras, point clouds
• Robotics, warehouse automation, or industrial perception experience

Benefits

• Ownership with a flat structure. No layers of approval between you and the people making decisions.
• You'll work alongside experienced engineers and researchers shipping autonomous systems into real warehouses, with plenty of room to grow your career as the company scales.
• Hybrid setup, and you shape your hours around your life. School pickup, dog walks, a lunchtime marathon training run, all fine.
• Flexible annual leave policy, employee stock ownership plan (conditions apply), a choice of pension plans, and relocation support if you're moving to Zurich.
• A dedicated buddy guides you through onboarding and your first month.
• Annual offsites, regular get-togethers, knowledge-sharing talks, and game nights. The office is dog-friendly, stocked with breakfast, fruit, snacks, drinks and great coffee.