ABOUT THE DIN

AI Detection Software Prototype for Electro-Optical ISR

InVision Tactical Group

About the Company

InVision Tactical Group is an Australian-owned Defence SME developing advanced electro-optical sensing technology for the Intelligence, Surveillance and Reconnaissance (ISR) domain. Its capability improves threat and target detection, reducing operator workload and speeding decision-making. InVision has validated its core capability through structured testing and is maturing the technology toward fielded products, with direct relevance to ADF ISR and land-domain priorities.

Project Objective

Automated AI detection directly supports the ADF by reducing the cognitive burden on soldiers and ISR operators who must otherwise scan imagery manually under time pressure. The intern will design and build, from the ground up, a documented software prototype that uses AI and computer vision (CV) to automatically detect objects of interest in imagery from InVision’s sensor system. The work spans establishing a secure development environment, building an image data and labelling pipeline, developing and training a detection model, and delivering a working or near-working prototype with complete build and process documentation — establishing the foundation for InVision’s sovereign detection capability.

Project Tasks

• Weeks 1–4: Establish a secure, isolated development environment and version control framework; ingest, organise and label InVision’s sensor imagery; build a repeatable data pipeline (import, pre-processing, labelling, dataset splits). Output: documented secure environment, working data pipeline, labelled dataset.
• Weeks 4–9: Select and implement model architecture(s); iteratively train and tune on the dataset; validate accuracy and false-alarm performance; document experiments, decisions and results; first milestone review with an AI subject-matter expert.
• Weeks 9–12: Integrate model into a working or near-working prototype with a simple inference interface; evaluate end-to-end performance; complete build, architecture and process documentation; final milestone review and handover pack for InVision’s continued development.

Intern Skills

• Python; solid grounding in computer vision and image processing (e.g. OpenCV).
• Machine learning / deep learning with PyTorch or TensorFlow, including CNNs for detection, classification or segmentation; experience building and labelling datasets and evaluating models (precision, recall, false-alarm rate).
• Disciplined software engineering: Git, reproducible pipelines, thorough technical documentation. Desirable: model integration into a working application; real-time or on-device inference experience.
• Self-directed; meticulous documentation habits; discreet and security-conscious; strong analytical mindset; clear written and verbal communication.
• Australian citizenship required; National Police Check, NDA and IP assignment to InVision required before starting. All work on InVision’s isolated workstation — no personal devices or cloud services. Willingness to be sponsored for future security clearance is desirable.

How to apply

Eligible students (Australian citizens) apply via the online application form and submit a CV and motivation letter by 6 September 2026.