At Holiday Tours & Travel, I designed and developed an interactive Power BI dashboard tracking three KPIs for more than 10 team leaders. It replaced monthly manual reporting with daily performance monitoring.
I also identified recurring issues in weekly profile error records and coordinated with account managers on root causes. By the end of the work, weekly records had decreased from 300–600 to 200–400; I also checked and resolved profile discrepancies using global data quality tools.
For my final year research, I built and evaluated Python machine learning models to predict 16-class MBTI from music preferences. Feature engineering on Spotify audio features helped me reach 68.2% accuracy, compared with baseline performance of about 40%.
My projects include a ConvNeXt-Tiny fish classification model and a prototype web app for real-time predictions from uploaded images. At the JomHack Varsity Hackathon, I developed a Python-based chatbot integrated with Telegram to help underprivileged and refugee schools find relevant NGOs, donors, and sponsors.

