John Lins

I study computer engineering (CpE) at the University of California, Santa Barbara.

My research journey began at Lawrence Livermore National Laboratory where I investigated AI-based and biologically-inspired optimization techniques for optimizing hydrodynamic simulations (ICLR AI4DiffEqn in Science 2024; Vienna, Austria). All my current and future research can be found on my Google Scholar.

My current research is being performed in dual collaboration with the CMU Language Technologies Institute and the UCSB NLP Group. We are investigating ways to map named entities in code to reduced token representations, this will enable LLMs to improve its understanding of code.

Additionally, I am the founder of Loyae.com, which aims to unleash the power of natural language processing (NLP) and machine learning for website optimization. Loyae intends to make the internet more accessible, searchable, and useable. The disgnostic tool has 1000+ users and the CMS plugin was recently approved to the WordPress plugin directory.

Lastly, I am working on a new language model for generating high-dimentional tensors from natural language prompts. I am using my model to offer Nparam Bull 0.1, software to predict asset prices from natural language prompts (of semi-public, non-priced-in information).