Work Experience
KLASS
Software Engineer (Speech Analytics)
May. 2023 - Present
- Built out full-stack Deep Learning and AI web applications that service homeland security, in particular the speech analytics domain.
- Productionised research output conducted on various aspects of speech analytics (antispoofing, speech diarisation).
- Collaborated with partners from Ministry of Home Affairs (MHA), Changi Airport Group, as well as other scientists from academia and industry to expand use cases in applied Deep Learning and AI research.
Python
Docker
Flask
React
Git
Tensorflow
Whisper AI
ZeroMQ
DSO
Member of Technical Staff
Aug. 2021 - Apr. 2023
- Sense-make human behaviour at various levels-of-analysis: individual cognition and decision making; small group dynamics; and macro-level social relationships
- Investigated abuse and threat actors on Social Media sites to understand techniques, impact
- Lead a cross-functional team of 4 to research and develop a rules based narrative recommender system based on social media text data.
Tableau
Python
Git
SPSS
selenium
R
Zk Analytics
Mixed Methods Researcher
Mar. 2021 - Aug. 2021
- Completed ad-hoc data analysis and data pre-processing tasks for a behavioural data science start-up
- Projects include: developing qualitative coding frameworks to access the mental health research for ODI, and conducting longitudinal survey analyses in R for WHO
R
MaxQDA
Research Areas
- During my time at the ANU and LSE, a substantial part of my work looked at the concept of a coalition president, and the patterns of coalition formation and maintenance. My graduate thesis studied the coalition dynamics and personal circumstances behind the formation of President Jokowi's day one coalition, finding that despite the ability of the Indonesian President to tap into various executive toolkits, they do not sufficiently explain the President's use of additional safeguards in his coalition management toolkit.
- During my research stints at Control Risks and AKE, I was mainly focused on Asia Pacific issues, and wrote several Op-eds and press summaries covering intelligence updates in the region.
- During my time as a behavioural data scientist, I created a human-in-the-loop active learning model that iterated on a multilabel dataset of social media textual data classified by different psychological traits. The news sources where these data were gathered was subsequently used to study media polarisation in social media, as well as hate speech detection based on different topic clusters.
Asian politics
Asian politics
Media polarisation and hate speech detection
Media polarisation and hate speech detection
Public Engagement
As mentioned, I continue to tinker and learn things in my free time, from being an enrolled at Georgia Tech’s online Masters of Computer Science programme, to learning some IoT and keeping tabs with the latest Web3 trends. As a result, you can see me at the IoT Meetup group at Singapore’s Metacamp events.