Simulation that lets clinicians practise, repeatedly.

The challenge

Medical training is limited by how often a procedure can safely be practised. Simulation removes that limit, but only if the experience is realistic enough to build genuine skill and consistent enough that a score means something from one session to the next.

Our approach

The work combined interface design with the engineering behind the simulation. We designed the interaction so the trainee's attention stays on the procedure rather than the software, and built the scoring loop — input, simulation, assessment — so results are comparable across sessions and trainees. Machine learning drives the assessment step, which meant treating the training data pipeline as a first-class part of the system.

The outcome

The result is a training tool clinicians can use repeatedly, with measurable output instructors can review. Because the assessment loop is a defined component, new procedures can be added without redesigning how scoring works.