Haldorix, a forward‑thinking startup studio, transforms operational challenges into scalable business solutions. We are seeking a Computer Vision/Machine Learning Engineer – Inference & Model Deployment to spearhead the development of a real‑time vision system that monitors productivity in textile production lines. Your responsibilities will include:

• Model Optimization & Deployment: Export deep learning models to ONNX, convert to TensorRT (INT8, FP16), and fine‑tune for efficient inference on GPUs.
• Active Learning & Data Pipeline: Manage data collection, annotation, and correction workflows using CVAT, ensuring continuous model improvement.
• Performance Benchmarking: Conduct GPU stress tests, measure latency, throughput, and accuracy across deployment environments, and document results.
• Field Validation & QA: Lead on‑site validation sessions, perform detailed QA checks, and guarantee robustness under real‑time production conditions.
• Automation & Tooling: Build scripts and tools to streamline model conversion, quantization, and deployment pipelines, integrating with Docker/Kubernetes for scalable deployment.
• Collaboration & Documentation: Work closely with computer vision, backend, and QA teams to integrate optimized inference pipelines and maintain comprehensive documentation.
• Continuous Improvement: Analyze bottlenecks, propose optimizations, and refine the active learning loop to enhance model accuracy.

Requirements include 2+ years of hands‑on experience in ML model deployment, strong expertise with TensorRT, ONNX, GPU optimization, and Python. Familiarity with CVAT, Docker, Kubernetes, CI/CD, and industrial computer vision is highly desirable. A BAC +5 degree and 3–5 years of relevant experience are expected. The role offers a hybrid working model in Casablanca, a collaborative culture, and the chance to shape cutting‑edge industrial vision solutions.