Working prototype
Edge AI Stethoscope
Separates heart and lung sounds, filters noise and flags patterns worth a second listen — entirely on the device, with no internet.
Devices & work
Four devices, one goal: good screening anywhere.
Each device is labelled with its real stage. Below them is the research they are built on and the earlier work that led here.
Devices
In development
Cough Listener
A small box for clinic waiting rooms that picks out coughs from background noise and suggests who may need TB or pneumonia screening.
In development
Single-lead ECG Patch
A thin chest patch worn for several days that watches for irregular heart rhythms such as atrial fibrillation and saves only the moments that matter.
In development
School Hearing Screener
A headphone kit that lets a school nurse run a quick hearing check, with a built-in microphone that pauses the test when the room is too loud.
Research foundation
SPMamba — 2-source to 3-source separation
Extended SPMamba from two-source to three-source speech separation, presented at UEC ASEAN Seminar and Workshop 2026.
- SPMamba
- State-space models
- Speech separation
MossFormer 2 — 3-source separation
Trained MossFormer 2 for three-source audio separation as a comparison point for SPMamba.
- MossFormer 2
- Transformers
- Audio separation
Earlier work
Face Recognition → Arduino
An early project connecting a camera, a recognition model and an Arduino output — the first step from AI software to physical hardware.
- Computer vision
- Arduino
Cisco Networking Labs
Structured Cisco Packet Tracer labs covering routing, switching, OSPF, EIGRP and NAT.
- Cisco Packet Tracer
- OSPF
- EIGRP
- NAT