Stuff

Stuff I did

Updated

Experience


National Serviceman, Singapore Army, until February 2027

  • Full-time active service in the Singapore Armed Forces.

Part time research intern, DSO National Laboratories, 2024 — Present

  • Working on dynamic compression for frontier sequence modeling architectures.
  • Played with a bunch of H200s lol

Education


National Junior College, High School Student, 2023-2024

Technical Skills


  • Programming Language: Python, C/C++, CUDA, MATLAB, Rust
  • Machine Learning and Deep Learning: PyTorch, JAX, MLSys Frameworks (CUTLASS, CuteDSL, Triton, TileLang, ThunderKittens), Statistical Methods, Optimization Methods
  • Office and Productivity Suites: LaTeX, Markdown, Doxygen

Selected Awards


Research

Regeneron International Science and Engineering Fair (ISEF): 3rd Place Grand Award for Robotics and Machine Learning, Association for the Advancement of Artificial Intelligence (AAAI) 1st Place Award, National Security Agency (NSA) 3rd Place Award; awarded May 2024, team project in Singapore with the highest number of awards at ISEF

Selected Work Experience


Research

Self-directed research: Student

Spedrun this for fun lol. Worked on improving the efficiency, accuracy, and expressitivity of newly released (at that point in time) state-space model (SSM) architectures like Mamba, S5, and Liquid-S4. Project selected to represent Singapore at ISEF.

  • Wrote a custom associative scan kernel in JAX and Triton to speed-up end-to-end training.
  • Proved that our architecture can express problems outside TC0\mathsf{TC}^{0}, including state-tracking problems like S5S^{5}.

Invited Talks


AAAI-25: Presented work at AAAI-25 Undergraduate Consortium.