Cheo Le Xian

Experience

• 

  AI Singapore

  Jan 2024 - Jan 2025

  AI Undergraduate Researcher

  - URECA program under NTU. - Conducted research on AI topic (speaker diarization), in speech and language domain. Focusing on developing state-of-the-art machine learning models to accurately identify and differentiate speakers in complex audio environment. - Improved overlapped speech detection on Dihard3 dataset from 69% to 87% accuracy.

• 

  DSO National Laboratories

  Jan 2025 - now

  AI Engineer

  - Working on Vision‑LLM, integrating open‑source image/video backbones with large language models to enhance multimodal understanding.- Mastered core VLM structures, end‑to‑end prompt engineering, data crawling methodologies, and fine‑tuning best practices through hands‑on research and continuous performance analysis.