Andrew Koh

I am a Machine Learning Engineer and Researcher with a Ph.D. in Computer Science from Nanyang Technological University, Singapore, specializing in audio and language understanding. While my primary focus has been leveraging deep learning to build cross-modal systems for audio captioning, retrieval, and acoustic scene analysis, I maintain a broad interest in the wider AI landscape, staying current with research in computer vision, generative modeling, and reinforcement learning to bring a versatile perspective to my work.

Alongside research, I enjoy building applied AI products end-to-end. I also take on independent and freelance engineering work, including web development and automation tools for small teams and businesses. I enjoy the challenge of identifying a gap and building the end-to-end product needed to fill it, making sure the final tool is as useful as it is technically sound.

Projects

Blog

• Building SakuraSensei: Notes From a Japanese Learning Experiment
  What started as a joke about my friend became a lesson in product thinking, user behavior, and the quiet satisfaction of finishing something
• What my Second PhD Paper Taught Me
  A personal reflection on an APSIPA paper, focusing on the pressures, constraints, and lessons that shaped the work beyond what appeared in the final publication.
• Building FaceChangerGIFBot: Reflections on Shipping a Side Project
  What started as a joke about my friend became a lesson in product thinking, user behavior, and the quiet satisfaction of finishing something
• The Hidden Struggles of My First Doctoral Paper
  A reflective look back at my first published paper at ICASSP 2022. The post moves beyond technical results to discuss the realities of PhD life: GPU shortages, the shift from architecture design to empirical iteration, and the personal motivation behind researching audio tools as a hearing-impaired individual.

Publications

Full list on Google Scholar.

Other Work