I build systems that perceive and reason across modalities — encoding, aligning and fusing text, image and sound. Associate Software Engineer at Razer, CS graduate from NUS.
Razer Inc.
Contributing to the development of Razer Ava, an AI personal assistant with real-time voice and video chat experiences. Built features for streaming voice, animation, and text transcripts, while working on desktop integrations for hardware discovery, wake-word model tuning, acoustic echo cancellation, and voice activity detection.
AI Singapore
Built a foundation across the full machine learning life cycle, including model architectures, knowledge representation, and MLOps. Later developed an automated compliance reporting system for sustainable food supply chains, combining data cleaning, statistical analysis, AI agents, and custom tools to support data analysis and report generation.
A*STAR
Curated and refined datasets from long, structured documents. Investigating data processing pipeline challenges and exploring cutting-edge Multi-Modal Large Language Models.
NUS Computing
Researched and developed a typescript API for React Native to interact with the iCOquit Smokerlyzer, a carbon monoxide sensor used in smoking cessation applications to measure carbon monoxide levels. Integrated the developed API into a smoking cessation mobile app using Expo React Native(TypeScript) and Firebase (Authentication and Database)
CSIT
Designed, implemented and deployed the backend of a collaborative meeting management system using SpringBoot, VueJS, ElasticSearch and NLTK within a 5-month timeframe, enhancing efficiency in meeting processes.
An empathetic companion that helps users identify burnout, journal their gratitude, and reclaim balance through conversational therapy.
A volunteering platform with robust volunteering management features for NUS students to sign up for volunteering opportunities and management from student clubs and organizations to manage volunteering events.
A meal planning application focused on promoting healthier lifestyles by providing personalized meal and diet recommendations tailored to user preferences.
An innovative running application that leverages real-time pace data to intelligently predict and signal users when it's time to head back.