"If your message matters, the name matters too."
We've all heard names mispronounced over loudspeaker systems in airports, hospitals, and other busy public-facing environments - followed by the same call repeated again to try and get it right. Technology and processes have come a long way, but spoken announcements are still largely manual, still common, and still prone to this kind of breakdown - and it's a global problem.
By combining machine learning with our own IP and modern text-to-speech systems, we've built a tool that makes these announcements easier and more understandable. It can handle any international names and can generate translated variants of the announcement when needed, so the correct customer hears the message the first time.
Founder + Full Stack Engineer
Experienced engineer across multiple sectors
Likes solving challenging problems with modern technology
Honours Degree in Engineering
After the initial shock that so many people were manually pronouncing names around the world, work began on the how to solve the problem
After much R&D, a complete project reset, and 100s of hours of machine learning, our first model and audio pipeline was ready
Progress was made on building a secure and functional REST API & cloud platform
The front-end of the product took shape and the platform's performance was evaluated
The model and logic was refined and re-trained on improved datasets
The API and secure gateway logic were further refined and secured
"Our names are all important, so let's use technology to help keep them that way."