Machnie Learning Engineer (CV model)

About the Role
We are looking for a Junior to Mid-Level ML Engineer to join the Model Engineering team. This role is suited for someone who enjoys working hands-on with computer vision models and understands models as systems to be trained, evaluated, improved, and optimised. You will work across the model lifecycle, including dataset preparation, architecture development, model training, evaluation, performance analysis, and optimisation for real-world computer vision applications.

Responsibilities

• Train, fine-tune, and evaluate computer vision models for classification, object detection, segmentation, pose estimation, and action recognition.
• Improve model performance through architecture enhancements, custom heads, backbone modifications, loss functions, and training strategy optimisation.
• Design experiments, conduct ablation studies, benchmark new techniques, and analyse model trade-offs.
• Investigate edge cases, debug failure cases, and improve model accuracy, robustness, and efficiency.
• Optimise models based on practical constraints such as latency, throughput, resource usage, and hardware limitations.
• Collaborate with engineering teams to improve dataset quality, model performance, and system reliability.

Requirements

• Bachelor’s degree in Computer Science, Artificial Intelligence, Machine Learning, Data Science, or a related technical field.
• At least 2 years of hands-on experience training, fine-tuning, and evaluating computer vision models.
• Strong hands-on experience with PyTorch and/or TensorFlow.
• Proficiency in Python and familiarity with Linux environments and command-line tools.
• Experience with computer vision tasks such as classification, object detection, segmentation, pose estimation, or action recognition.
• Experience modifying models beyond configuration-level changes, including architectures, backbones, heads, loss functions, or training pipelines.
• Experience designing experiments, conducting ablation studies, analysing model performance, and debugging failure cases.
• Familiarity with multimodal, vision-language, foundation, or open-vocabulary models.
• Experience with model optimisation techniques such as quantisation, pruning, or distillation.
• Familiarity with experiment tracking, reproducibility, and structured machine learning workflows.