Stephen Z. Lu


Hey there, welcome to my blog! My name is Stephen Lu and I am a third year undergraduate student in computer science and biology at McGill University.

My interests are quite broad, but I am particularly fascinated by the areas of graph representation learning, NLP, and generative modelling with applications in chemistry and biology. The microscopic world has always fascinated me — how is it that from a handful of atoms and nucleotides, when properly arranged, emerges such a rich diversity of materials and life forms? How can we better understand this relationship between the low-level code of science (atoms, bonds, DNA, etc.) and the high-level properties that characterize their assemblage into useful materials, drugs, and nutrients?

Unfortunately, probing the microscopic world for the answers to such questions remains an arduous endeavour due to the combinatorial explosion of possible interactions and arrangements that plague atomic systems of increasing complexity (this is partly why experimentalists have my utmost respect). However, if we make a few reasonable assumptions, then the problem becomes less daunting, at least for computers (stay tuned for a new blog post where I’ll address this in detail). In this light, my primary research interest lies in developing new representation learning techniques that can adequately model the low-level code of molecules and genes.

In my free time, I love playing pickup basketball at my local park (Valois park in Pointe-Claire if you’re from the area), composing mostly funky (and a few serious) songs on the piano, discovering breathtaking landscapes nestled in low-profile hiking trails (my profile pic above was taken atop Mt. Wakakusa in Nara, Japan), and spending time with my family and friends in my hometown of Montreal, Quebec.

I was previously a co-founder and lead engineer at Hero AI, where I designed and developed the core frontend architecture for the flagship real-time dashboard and patient beacon products, did research on clinical screening models at the Hospital for Sick Children, and completed internships at Lino AI, The Bagot Lab, and The Ferbeyre Lab.

I am also very grateful to be a Schulich Scholar and an alumnus of the Canada-Wide Science Fair where I started my STEM journey.


Dec 27, 2023 [new blog post!] Latent Dirichlet Allocation (LDA)
Nov 19, 2023 [new blog post!] Degenerate Dot Product Attention
Oct 28, 2023 [new blog post!] Variational Inference Implemented (Part 2)
Oct 9, 2023 [new blog post!] Variational Inference w/ EM algorithm (Part 1)
Sep 24, 2023 [new blog post!] LLM Finetuning w/ SMILES-BERT