Alongside his technical role as an engineer—spanning regenerative design, natural building materials, and the reuse of existing buildings—Mike has built a career-long interest in computational thinking and its practical application in project delivery. He has contributed to internal computational groups of engineers and software developers to create parametric approaches that support complex structural design while improving consistency and unlocking faster iteration.
More recently, Mike has deepened his focus on data and AI, completing formal training covering data manipulation, feature engineering, supervised and unsupervised learning, data visualisation and communication, and developing and training large language models. This connects directly to a wider industry shift towards better monitoring, recording, and use of information to better understand our buildings and how they interact with the natural environment.
For Mike, when AI is used well, it can support earlier insights, better optioneering, clearer communication, and more time spent on the engineering decisions that matter. Used poorly, they risk driving everyone toward the same patterns, the same blind spots, and the same outcomes, especially if teams don’t actively test, validate, and challenge what the tools produce.
“The problem we have in the automation landscape at the moment is that if all companies decide to use the same models, asking them the same questions, then we get the same outcomes as everyone else. This workflow is missing your unique identity. Your previous work is a catalogue of your culture and your identity. Use it to train models in the way that you think and you work. “
“We need better understanding. We need people, in academia and industry, researching these tools and how they could be used more efficiently, discovering their capabilities and drawbacks, and discussing when they are most appropriate.”