We build robust computer vision models capable of interpreting complex visual environments, ranging from microscopic details to large-scale aerial landscapes. This includes deep learning for object classification, temporal and spatial change detection in high-resolution imagery, shadow removal, and automated semantic segmentation. We actively apply these competencies to enhance national mapping infrastructure and optimise public services, which you can read about in our Projects Section. To see the academic frameworks behind these models, you can browse our peer-reviewed papers in our Publications Section.
We specialize in integrating intelligent software frameworks with physical hardware to solve operational challenges in difficult-to-access terrains. Our work covers AI-guided drone flight path planning, multi-stage small object detection from variable altitudes, and autonomous decision-making algorithms for robotic manipulation. These initiatives directly support national environmental conservation efforts, which we outline in detail in our Impact Section. To safely execute these missions, members of our multidisciplinary team - whom you can meet on our People Page - hold certified European drone licenses for legal and compliant aerial data acquisition.
Our team engineers advanced language models and multimodal systems designed to parse, categorise, and track information flows within dense digital media ecosystems. We focus on named entity extraction, multi-source content segmentation (spanning image, video, and audio classification), sentiment analysis, and conversational information exploration utilising Large Language Models (LLMs). These systems form the foundation of several user-facing platforms in our Projects Section that help journalists, regulators, and citizens analyse broadcast and digital media structures objectively.
We do not just build "black-box" models; our research emphasises algorithmic transparency to identify, track, and measure societal trends and structural patterns. Our technical competencies include developing interpretable machine learning models, creating visual bias tracking frameworks, and structuring automated tools for media representation analysis. This focus on equity, transparency, and institutional accountability directly maps to our commitment to the United Nations Sustainable Development Goals, fully detailed in our Impact Section.
We extend our technical expertise into the public sphere, creating the educational and strategic frameworks required to sustain a technology-driven economy. We specialise in instructional design for diverse demographics, ethical AI framework development, modular learning pathways, and public engagement strategies. Our landmark work in this area includes designing national-scale baseline learning programs found in our Projects Section. These massive public-facing campaigns are executed in tandem with prominent national entities, whom you can view on our Collaborators Page.
