David Austin

PhD student at MILA/McGill

Career

I graduated from the University of Waterloo in April 2023 with a major in Computer Science and minor in Statistics. I’ve completed several internships working with machine learning (ML) and natural language processing (NLP). I hope to pursue further studies in the field.

My research interests include the following:

  • natural language processing
  • intelligent interactive systems
  • multimodality
  • information retrieval (particularly recommendation systems)
  • human-computer interaction
  • social impact of AI

Previous Work

In 2023-24, I was a faculty affiliate researcher at the Vector Institute in Toronto, with Dr. Scott Sanner and Dr. Chris McIntosh of the University of Toronto.

I led a project with Dr. Sanner to extend decision-theoretic conversational recommendation systems to handle an unbounded language space. We combine the explicit models provided by decision theory with the free language interaction of LLMs.

I worked with Dr. Chris McIntosh’s lab to develop a multimodal foundation model that is able to unify medical modalities (e.g. ECG, CXR) using text as a shared embedding space (similar to Med-PaLM-M and ImageBind).

See my CV for work during my undergraduate studies.

Hobbies

Outside of my research work, I like to play music, read, run, draw, and more.

This year (2023), I ran my first ever 10K with a time of 0:50:28 and played my first ever live musical performance (on guitar and bass).