Welcome! I am currently a master’s candidate at SEAS in Harvard University.
(Let’s get in touch! I’m best reached through my email, \< albertge /> at g.harvard.edu.)
I am a research fellow at DASlab, advised by Prof. Stratos Idreos, where I think about how to design neural network architectures. I have particular interests in network compression, to reduce training costs and speedup inference time. I also work with Prof. Cengiz Pehlevan, to characterize scaling laws of data and model performance. I’m exploring ideas from computational neuroscience to inform how we can achieve better generalization with fewer examples.
Broadly, I’d like to understand these questions:
- What is the design space of efficient neural network models?
- Can we effectively incorporate geometric knowledge from other scientific theories into our artifical models?
Before deciding to get into research, I was a full-stack software engineer at Academia.edu and at Abbvie Stemcentrx. I obtained my bachelor’s degree in computer science from Caltech in sunny Pasadena.