Research Assistant at Nanyang Technical University - NIE Singapore (2026-02 – Present)
Conducted affective research to develop Open Emotional Learner Models and engineered privacy-centric deep learning pipelines.
- Conducted affective research to develop "Open Emotional Learner Models" that map facial dynamics to student learning states for improving student learning.
- Engineered a privacy-centric deep learning pipeline using the DAiSEE dataset, extracting facial blendshapes to train accurate models without storing raw video.
Product Analyst at Powerup Money (2025-09 – 2025-12)
Formulated and tested data-driven hypotheses on user behavior and product engagement.
- Formulated and tested data-driven hypotheses on user behavior, evaluating how product changes influenced user engagement and progression through key interaction stages.
Product Analyst at PocketFM (2025-09 – Present)
Developed pLTV models, conducted behavioral analysis of high-value users, and managed production-scale data pipelines.
- Developed and deployed a pLTV model to forecast Day-30 user revenue, and built the end-to-end data pipeline supporting the prediction workflow.
- Conducted in-depth behavioral analysis of high-value ("Whale") users to uncover early predictive signals, modeling temporal user interaction patterns and translating them into features for monetization and retention prediction tasks.
- Managed and maintained production-scale user data pipelines and fact tables, ensuring data quality and consistency to support accurate, data-driven decisions by business stakeholders.
Analyst Intern at PocketFM (2024-01 – 2024-06)
Contributed to data infrastructure migration from Amazon Redshift to Databricks.
- Contributed to the migration of data infrastructure from Amazon Redshift to Databricks, improving processing efficiency, scalability, and reliability of analytics workflows.
Research and Algorithm Developer at Happy Reliable Surgeries (2023-06 – 2023-07)
Registered MRI scans and experimented with AI-based medical image segmentation methods.
- Registered MRI scans of the spine and brain with 3D camera data using the Iterative Closest Point (ICP) algorithm, improving spatial alignment for medical imaging analysis.
- Utilized segmentation tools such as 3D Slicer and ImFusion to preprocess MRI images and experimented with AI-based methods for automating medical image segmentation.
