Interview: Alessandra Galli, James Dyson Award 2025 Medical Winner

I’ve always known about the James Dyson Award because it’s a really well-known competition for students. Everyone really looks up to this competition because the projects are really relevant and interesting – I could always learn something from them. So it has always been a dream of mine to apply, and I felt that with OnCue I’m solving a really important problem that is often overlooked, so I finally did. 

Winning the James Dyson Award is both an honour and a confirmation that committing to this project after graduation was the right decision. The James Dyson Award offers not only recognition, but also a practical opportunity to move OnCue forward. The prize provides significant support for completing the operational prototype, bringing the project one step closer to reaching people with Parkinson’s. I am confident that this achievement will also help me expand my network and open new opportunities for OnCue.

The Al in the keyboard works by analysing the letters users type and predicting which ones are most likely to come next, based on language patterns and context. Users can physically select on the keyboard whether it should operate in their native language or in English. When a key is pressed, it lights up in yellow for instant feedback, while the most probable next letters are illuminated in green.

Technically, the system is powered by a predictive language model trained on text datasets. It uses probabilistic algorithms to calculate the likelihood of each possible next letter, given the sequence of characters already typed. This allows the Al to adapt dynamically to different languages and individual writing patterns.

As the user continues typing, the model refines its predictions in real time. The more letters are entered, the fewer keys remain highlighted, as the system becomes more confident about the next character. This close interaction between the Al model and the LED control system creates a continuous feedback loop between cognitive intent, visual guidance, and motor response.

Ultimately, this process not only helps guide the user and improve typing fluency, but also reduces moments of freezing or hesitation by maintaining a smooth and uninterrupted flow, which, during user trials, emerged as their top priority.

They are enthusiastic that there’s something especially designed for them. There are assistive keyboards on the market right now, but there isn’t one that’s specifically designed with Parkinson’s symptoms in mind.

I’ve received a lot of positive feedback for the haptic wristbands. People described that they give a feeling of calmness and relaxation. They also said that the vibrations in the keyboard help with their awareness when pressing a key. 

Ensuring that the product can easily reach its intended users is a top priority. OnCue is not designed to be a niche product, but a widely accessible one – which is why affordability will remain a central focus throughout its development.

To achieve this, alongside refining the design and technical details, I’m also researching how people with Parkinson’s in different countries access assistive technologies to better understand potential barriers and opportunities. These insights will help me define strategies to ensure that OnCue remains economically accessible once it reaches the market.

Day by day, the OnCue team and network are growing. OnCue is now endorsed by ParkinsonNL and Confederazione Parkinson Italia. In the next few months, we will focus focus on refining the electronic and mechanical components of the product, leading to the development of fully functional, full-size prototypes.

These prototypes will undergo one month of structured usability testing in collaboration with occupational therapists and private stakeholders in the Netherlands and in Italy, through the network of the Parkinson’s Association of Treviso. We will work with 10–12 people living with early-stage Parkinson’s, who are self-sufficient and regularly use computers. This pilot phase will provide real-world validation of the device.