Winner

Red Hat Hackathon

September 2017

Computer-Vision based solution to track crowds

PYTHON OPENCV IOT
Prototyped the usage of a Raspberry Pi and a ceiling-mounted camera to track the movement of individual people in a crowd, to derive useful metrics such crowd size, movement and dwell time. This data, along with ambient noise information was then transferred via MQTT and made available via a Telegram bot meant for potential visitors to the area being monitored. This would allow them to better understand conditions there before attending the event.
GITHUB RED HAT

Winner

'Fake News Must Die' Hackathon

October 2017

Using Machine Learning to Combat 'Fake News'

PYTHON NLP JUPYTER NOTEBOOK FIREBASE JAVASCRIPT
Curated a data-set of online news articles pertaining to Singapore and made estimates on their credibility using Machine Learning techniques. Proposed an online social platform for users to collaboratively fact-check online news articles to improve the accuracy of the estimate. This will generate data valuable to researchers to develop new methods of weeding out 'Fake News'.
GITHUB STRAITS TIMES

Winner

#startathon 2017

 
October 2017

Autonomous Crane Inspection

ARDUINO ELECTRONICS PYTHON JAVASCRIPT

Created a magnetic, wall-climbing robot that can scale tall metallic structures vertically to perform inspections. This was to answer PSA’s challenge statement of performing autonomous crane inspection without using drones. The idea was selected by PSA for further development.