All case studies
AI & Healthcare
AI-Powered Diagnostic Tool Development
Collaborated with a Chinese health-tech firm to build an AI diagnostic prototype.
Example project. This describes the shape a CBIC collaboration takes and the work involved. Partner names, figures and dates are omitted rather than estimated — CBIC publishes those only with the partner's agreement.
The challenge
Diagnostic imaging expertise is concentrated in a handful of Pakistani cities, while the need is nationwide. A student team proposed a screening aid that could run on modest hardware in a district clinic and flag cases for specialist review — useful precisely because it does not attempt to replace the clinician.
How it worked
1The team submitted the project through their CBIC dashboard for expert screening.
2CBIC matched it to a Chinese health-technology partner working in medical imaging.
3Chinese engineers reviewed the model architecture and, critically, the training data strategy — the part student projects most often get wrong.
4The team rebuilt the pipeline around a smaller model that could run without a dedicated GPU.
5Clinical framing was tightened so outputs read as 'refer for review', never as a diagnosis.
Technologies involved
- Convolutional neural networks
- Medical image preprocessing
- Model quantisation for low-power hardware
- Python, PyTorch
What the team gained
- A working prototype demonstrated to the partner's technical team
- Mentorship from engineers who ship medical software commercially
- A project the team could defend in a final-year viva and show to an employer
- An ongoing technical relationship with the partner
If you are submitting a project
Screening favours projects that solve a specific, bounded problem. A narrow tool that works beats a broad platform that demos.
How FYP sponsorship worksHave a project like this?
Register once for a CBIC ID, then apply for this and every other member service from your dashboard.