Two students, two paths. Both started with projects.

Normal (top) vs. pneumonia (bottom): the chest X-ray data he worked with.
From physics to AI research in biotech.
Physics at NUS. No AI background.
7 projects across finance, NLP, computer vision and biotech.
AI + biotech. In a medical-imaging project, he found that simple data augmentation beat CycleGAN for pneumonia detection.
Read the projectAI researcher, Osaka University.
From computer science to health-tech AI.

A retina scan (left) and the blood vessels her model traced from it (right).
First-year computer science at NUS.
5 projects, plus two weeks on site in Cambridge.
Medical imaging. She built retinal vessel segmentation models that work with limited data.
Remote intern at a Boston health-tech startup, building computer-vision models for camera-based health monitoring.
The startup reviewed shortlisted profiles and chose her.
Labs and companies decide who they take. What you bring is work they can see.
Under a minute · Nothing to sign up for
Company names identify the industry context of project work.
Under a minute · Nothing to sign up for