Machine Learning Engineer - Contract

This contract offers the opportunity to take full technical ownership of a high-impact computer vision workstream focused on real-world AI deployment. You will lead the development and improvement of machine learning models that detect damage on large commercial vehicles using image data captured in operational environments. The work is practical, fast-moving and delivery-focused, with models moving into production rather than remaining in research. Ideal for an ML engineer who enjoys solving messy, real-world problems and driving measurable performance improvements.

Responsibilities Design, train and optimise computer vision models for vehicle damage detection using object detection and segmentation techniques Improve model precision, recall and overall accuracy across priority damage categories through structured evaluation and retraining cycles Work closely with data and annotation teams to define damage classes, identify data gaps and address class imbalance Carry out detailed error analysis to understand false positives and negatives and guide targeted model improvements Own evaluation datasets, testing methodology and performance reporting across training, validation and test sets Collaborate with platform and MLOps teams to package, deploy and monitor models in production environments Identify performance issues such as data drift and support rollout of inspection capability to additional operational sites

Strong commercial experience in machine learning with a focus on computer vision Hands-on experience training and deploying object detection or segmentation models such as YOLO or similar architectures Proficiency in Python and common ML and computer vision libraries Experience working with large image datasets and noisy real-world data Ability to translate operational or business problems into measurable ML objectives Comfortable working in an iterative, delivery-focused engineering environment

What's in it for me? Ownership of an end-to-end ML problem with visible operational impact Opportunity to build and scale production computer vision systems Close collaboration with data, operations and engineering teams Exposure to complex, real-world datasets and deployment challenges A clearly defined problem space with autonomy and accountability