Job Description

Posted 5 days ago

Data Scientist

Are you passionate about pushing the boundaries of Machine Vision? We’re looking for a talented Data Scientist specialising in Machine Vision or Spatial Analysis to develop, train, and deploy advanced image and video detection models across EO, IR, near-IR, and multi-band data sources.

You’ll work at the forefront of deep learning and classical computer vision, designing solutions for challenging real-world applications including small and distant object detection, multi-modal fusion, and real-time performance optimisation.

Key Responsibilities

*

Design, train, and optimise detection and classification models using deep learning and classical computer vision techniques.

*

Manage and prepare large-scale imagery datasets, including labelling, augmentation, and quality checks.

*

Develop algorithms for multi-spectral and fused EO/IR/near-IR data.

*

Validate, benchmark, and refine models for deployment in real-time or embedded environments.

*

Collaborate with engineering teams to integrate, deploy, and maintain production-grade vision models.

Requirements

*

Strong expertise in machine vision, including CNNs, transformer-based architectures, and classical CV methods (OpenCV, scikit-image).

*

Proficiency in Python and key ML/vision libraries (PyTorch, TensorFlow, NumPy, SciPy, scikit-learn).

*

Experience designing models for detection/classification (YOLO, Faster R-CNN, RetinaNet, ViT, EfficientNet).

*

Hands-on experience with imagery across EO, IR, and/or multi-band datasets.

*

Solid understanding of machine learning fundamentals, training workflows, and evaluation metrics.

*

Master’s/Ph.D. in Computer Vision, Machine Learning, Data Science, Applied Mathematics, or related field, plus 2–5 years’ relevant experience.

Soft Skills

*

Strong analytical mindset with exceptional problem-solving abilities.

*

Clear communication and documentation skills.

*

Proactive, collaborative, and comfortable working in a fast-paced, research-driven environment