Qualifications
ResponsibilitiesOversee the development and deployment of computer vision models (including detection, tracking, and multimodal vision). Guarantee technical excellence throughout the entire model lifecycle: data acquisition, training, evaluation, and cloud-edge deployment. Architect and refine the technical framework of computer vision systems (vision stack, ML infrastructure, and pipelines). Foster the growth of team members through coaching, mentoring, and feedback sessions. Collaborate closely with the Head of AI and product teams to align technical strategies with business objectives. Implement and uphold best practices concerning code quality, ML CI/CD, MLOps, monitoring, and documentation. Engage actively in technology selections and remain informed about advancements in applied computer vision and AI. ProfileKey Skills (Essential)In-depth expertise in applied computer vision with hands-on experience in production settings. Proficient in Python and principles of software engineering best practices. Advanced knowledge of PyTorch (or TensorFlow). Solid experience with detection and tracking models (e.g., YOLO, R-CNN, DeepSORT). Strong foundation in MLOps: including pipelines, monitoring, versioning, and deployment. Capability to manage and resolve complex technical challenges effectively.
About the job
Mission
At Vizzia, you'll join a dynamic deeptech startup dedicated to developing cutting-edge solutions for urban environments.
The AI division is currently focused on two primary products:
Waste: An automated system for detecting illegal dumping, which is in the process of scaling.
Safety: A state-of-the-art urban monitoring solution that is being developed from the ground up.
As the Engineering Manager – Computer Vision, you will have a pivotal role:
Developing robust, production-ready computer vision systems,
Leading and nurturing a team of engineers and data scientists in a challenging, real-world product setting.
Your contributions will directly influence the quality, reliability, and effective deployment of models in a live production environment.