About the Role
As a Computer Vision Engineer, you will lead the development and continuous improvement of our state-of-the-art AI models powering the FleetVision platform. You will take end-to-end ownership of model development, translating clear business and safety requirements into robust, production-grade computer vision solutions using cutting-edge technology. You will play a key role in shaping our computer vision roadmap, improving detection accuracy, reliability, and scalability across large-scale maritime CCTV deployments.
Key Responsibilities
- Lead the end-to-end development lifecycle of computer vision solutions, from problem definition and data collection through research, training, evaluation, deployment, and ongoing monitoring.
- Design, implement, and iterate on advanced computer vision and deep learning models tailored to real-world maritime environments.
- Work closely with cross-functional teams, including product, data, and platform engineers, to deliver impactful, production-ready solutions to complex maritime safety and operational challenges.
- Own model quality in production, including validation strategies, performance analysis, failure investigation, and continuous improvement.
- Stay current with relevant research and translate state-of-the-art methods into practical, scalable systems.
Qualifications
- 7+ years of hands-on experience in Computer Vision and/or Deep Learning, with a proven track record of taking models from POC to production.
- Experience working with CCTV or video-based datasets is required.
- Master’s degree in Computer Science, Engineering, or a related technical field.
- Strong proficiency in Python, with the ability to write clean, maintainable, and scalable code for large, long-lived projects.
- Deep expertise in PyTorch and solid experience with OpenCV and other computer vision libraries.
- Strong mathematical foundation and the ability to read, understand, and critically evaluate research papers and algorithms.
- Bonus if you have experience deploying and optimizing models on edge devices or resource-constrained environments.