Hok Seng

Engineering degree in applied mathematics and artificial intelligence

Coursework: Computer Vision, Natural Language Processing, Machine Learning, Deep Learning

Two years of core curriculum: Mathematics, Physics, Computer Science

Implemented a Personalized PageRank algorithm for multi-hop questions and designed an agent system to answer a wide range of queries.

Implemented and optimized the GLiNER classification model with LoRA fine-tuning on a synthetic dataset generated via Gretel.ai; benchmarked against a BERT-BiLSTM-CRF baseline and spaCy (en_core_web_sm).

Improved the OMR pipeline using a YOLO model and a graph neural network to convert raw sheet music into MusicXML.

Built an end-to-end multi-agent travel chatbot as a team, using multimodal models to book hotels, recommend attractions, and generate invoices with scannable tickets through voice and chat interfaces, deployed on Cloud Run.

Developed data-imputation methods for ecological datasets with missing standard deviations, enabling meta-analyses with a multilevel model to estimate effect sizes and their variance. Won the 'Best Project' award from the Applied Mathematics Department.

Organize seminars, inviting experienced speakers to share knowledge on selected topics.