Machine Learning and AI
I am the Acting Team Leader and a Principal Research Scientist in the Data Vision Team. My expertise lies in the exciting fields of AI, machine learning, data science, image processing, and computer vision. Throughout my career, I have also worked extensively in telecommunications, big data, feature detection, information theory, and cognitive radio.
My passion lies in bringing technological innovation to businesses in New Zealand. I believe that leveraging the power of AI, machine learning, and data science can revolutionise how organisations operate and drive them towards greater success. One of my major influences is David MacKay, renowned for work on the value of extracting insights from data and the power of information analysis.
A career highlight was being involved in the design of the Square Kilometre Array (SKA), collaborating with researchers from around the world. Working on such a large-scale multinational R&D effort, where I explored the Ghost Phenomenon and astronomical transient data mining, reinforced the reality of humanity coming together to solve complex problems.
Currently, our Data Vision R&D explores small-world customised language models, trusted language models tailored to specific tasks. By focusing on the accuracy and specificity of models within defined domains, we prioritise precision over generality, ensuring superior performance in scoped topics. This contrasts with the training process for models like GPT-3 and ChatGPT which involved a supercomputer with approximately 10,000 GPUs, 285,000 CPU cores, and 400 Gbps of network connectivity for each GPU server. It still took 9 months to train and OpenAI/Microsoft spent USD $1billion on renting the hardware!