Ben is passionate about building AI that people can trust. His work focuses on developing machine learning models that are accurate, efficient and designed to perform reliably in the real world, helping ensure technology supports better decisions without adding unnecessary complexity.
Ben Winter
Machine Learning Engineer
Connect with BenBen is passionate about building AI that people can trust. His work focuses on developing machine learning models that are accurate, efficient and designed to perform reliably in the real world, helping ensure technology supports better decisions without adding unnecessary complexity.
With a PhD in Machine Learning from Bangor University, Ben brings a research-led approach to AI development, combining academic expertise with practical engineering. He specialises in natural language processing, automatic speech recognition and model optimisation, helping ensure Wyser’s technology is accurate, efficient and trusted in real-world environments.
At Wyser, Ben leads the development of the machine learning models that power WyserASSIST. From improving transcription accuracy across regional accents to reducing bias and increasing efficiency, he works closely with the wider team to ensure every model delivers meaningful value while remaining responsible, transparent and fit for purpose.
His work also supports Wyser’s ongoing research programme into responsible AI, helping improve the accuracy, fairness and performance of the technology that underpins advice and guidance services.
What drives me?
“What drives me is building a team that doesn’t fall in love with its first idea. The data usually turns out to be telling a different story than we thought at the start, and I want people who go looking for that rather than defending the plan.”
In three words:
Determined, Curious, Dependable.
What’s one misconception about AI you’d love to change?
There is so much more to AI than LLMs. In my PhD thesis I wrote “There is no need to kill a mosquito with a cannon”. Too many problems today get thrown at an LLM when a much smaller model would do the job, resulting in wasted money, resources and needless harm to the environment.
What excites you most about the future?
Model efficiency. The last five years have been about scale, and I believe the next five years will be about doing more with less, smaller models, cheaper inference and running things on your own hardware. I believe that will be a much more interesting problem to solve!
Outside of Wyser
I love sports, football, squash and padel particularly. I also enjoy writing programs and applications that help in my daily life… such as keeping my plants alive.