Haldorix is a dynamic startup studio that transforms operational challenges into scalable business opportunities through cutting‑edge technology. With a focus on industries such as manufacturing, retail, and logistics, the company develops dedicated solutions that evolve into standalone products. Two flagship ventures include DeVisu, a computer‑vision platform for operational optimisation, and Nitra3AI, a generative AI solution for industrial performance enhancement.

The Lead Computer Vision Engineer will spearhead the development of a real‑time vision system tailored for textile production monitoring. This system will capture high‑resolution video streams, perform geometric calibration to convert pixel coordinates into real‑world measurements, and apply advanced detection and tracking algorithms to identify defects and anomalies in the production line.

Key responsibilities include:

• Lead Computer Vision Pipeline: Design and manage the end‑to‑end video‑processing chain, from acquisition to detection, tracking, and visualization.
• Geometric Calibration: Compute homographies and other calibration parameters to translate pixel data into millimetre‑scale metrics, ensuring accurate measurement of textile dimensions.
• Model Optimization: Fine‑tune the FACT‑tiny model for high performance and robustness in real‑time inference, balancing accuracy with computational efficiency.
• System Reliability: Monitor mean average precision (mAP) metrics, detect model drift, and troubleshoot issues in live environments to maintain consistent performance.
• Documentation: Maintain structured finite state machine (FSM) diagrams, model specifications, and deployment guides to support knowledge transfer.
• Cross‑Functional Collaboration: Work closely with lead developers and product architects to align system components with product roadmaps and technical constraints.
• Performance Tuning: Optimize GPU usage, latency, and throughput for on‑edge or on‑premise deployments, ensuring the system meets real‑time processing requirements.
• Knowledge Transfer: Mentor internal teams, provide clear documentation, and facilitate hands‑on training sessions to build internal expertise.

Ideal candidates will possess:

• 3+ years of hands‑on experience in computer vision, including OpenCV, detection, and tracking.
• Strong understanding of camera calibration and 2D/3D projective geometry.
• Proficiency in Python and scientific libraries such as NumPy and OpenCV.
• Experience with deep learning architectures for video processing and GPU optimisation.
• Familiarity with edge deployment, lightweight inference frameworks, and DevOps workflows (Docker, Kubernetes, CI/CD).
• Prior work in industrial automation or IoT vision systems is a plus.
• Excellent documentation and communication skills in English or French.

The recruitment process includes a Jobzyn AI interview (25–45 minutes), a technical interview with the lead developer or technical architect, a practical test based on a real‑world use case, and a final interview with the team and partners.

Benefits of joining Haldorix include direct, measurable impact on industrial productivity, exposure to real‑world AI challenges with immediate operational deployment, and the opportunity to shape and deliver a full system end‑to‑end. The role offers a hybrid working model, competitive compensation, and the chance to work alongside a team of experts combining AI, software engineering, and industry know‑how.