• Led development of an AI-based product solution leveraging computer vision models to identify and remove violating items on the Shopee platform as part of Listing QC team
• Collaborated with cross-functional teams, including regional ops, UI designers, software engineers, and data science teams to ensure platform cleanliness
• Presented PRD to numerous stakeholders, articulating launch roadmap, technical specifications and performance metrics

• Designed and developed a full-stack computer vision data visualisation tool for Hyundai’s AI Research centre with Next.js, Express.js and MongoDB, utilising Docker for deployment on Kubernetes-based infrastructure
• Reduced API response data latency by 80% through server-side segmentation-mask decoding with LRU cache
• Adopted microservices architecture following clean architecture principles for integration across various CV/ML applications
• Conducted research on WINClip for zero-shot anomaly detection in Ioniq 5 Electric Vehicle manufacturing process

• Developed mobile interfaces using Flutter with Dart, integrating features such as profile page and earning analytics
• Migrated app database from Cloud Firestore to PostgreSQL with Hasura and GraphQL by mapping complex data relationships, resulting in 60% reduction in query response time
• Implemented robust BLoC stream-based state management system for future codebase maintainability and scalability

• Engineered a full-stack wellness mobile application to monitor users’ daily routines and predict wellness status via regression analysis, utilising Flutter SDK with Dart, Firebase and Heroku for backend hosting
• Leveraged Python and Scikit-learn to implement a DFS algorithm to identify associations and make predictions
• Conducted rigorous unit, widget, and integration tests ensuring quality and reliability of release build
• Awarded Judge’s Choice Award - out of 414 projects at highest level of achievement (Artemis) in NUS Orbital

• Developed a timetable optimizer for NUS students using a Genetic Algorithm to generate highly efficient and personalised schedules utilising JavaScript, HTML, CSS, Flask, Python, Selenium, and NUSMods API
• Integrated crossover and mutation operations to generate optimised timetables, accounting for constraints such as class availability and student preferences

Awarded out of 414 projects, earning the highest level of achievement - Artemis in NUS Orbital Programme. This award is given to the top teams selected by a panel of judges consisting of faculty members and IT professionals.

Award in recognition of outstanding academic performance

Singapore-Cambridge GCE A-Level