WritingScholarLinkedIn

Hello I'm

Jonathan Li

I’ve been messing around making what I
think is cool stuff—machine learning, design compilers—for the past eight (8) years.

who am i?

I mostly grew up in Ontario. But a vivid year that's shaped my interests more than any other was the year I lived abroad in Hangzhou, at a public school when I was 8.

That year I spent evenings exploring our family laptop. I stumbled upon my brother's pirated version of the Adobe suite, and spent the next year learning how to edit and composite VFX videos in Adobe After Effects with online tutorials. I distincitly remember hogging the laptop for hours at a time waiting on renders, to my siblings' great annoyance.

At the end of that year I made my first dollar on Fiverr by helping people edit videos. I still remember how thrilling it felt to make something other people were willing to pay for, as someone who had just taught himself how to do that thing.

Not long after, I learned to code and became captivated by the magic behind a neural network. The product of my exploration was published NLP research (back when it was still called that) including a first-author paper at ACL 2023. When I was 17, I joined Cohere as their youngest researcher, where I worked on reasoning, long-context, and reinforcement learning infrastructure. Now, I'm building a company with my brother.

I’ve recently started thinking a lot about my path, which has given me a lot of conviction about how we should build the next generation of ai.

Image of me

publications

prior to Cohere, I published a few papers
@ Queen’s U, with Prof. Xiaodan Zhu

Prefix Propagation: Parameter-Efficient Tuning for Long Sequences

Association of Computational Linguistics 2023 (ACL 2023)

Jonathan Li, Will Aitken, Rohan Bhambhoria, Xiaodan Zhu


Command A: An Enterprise-Ready Large Language Model

arXiv preprint (2025)

Cohere Team


Parameter-Efficient Legal Domain Adaptation

Natural Legal Language Processing Workshop 2022 (NLLP @ EMNLP 2022)

Jonathan Li, Rohan Bhambhoria, Xiaodan Zhu


Evaluating AI for Law: Bridging the Gap with Open-Source Solutions

arXiv preprint (2024)

Rohan Bhambhoria, Samuel Dahan, Jonathan Li, Xiaodan Zhu


Experimenting with Legal AI Solutions: The Case of Question-Answering for Access to Justice

Workshop on Generative AI and Law '24 (GenLaw @ ICML 2024)

Jonathan Li, Rohan Bhambhoria, Samuel Dahan, Xiaodan Zhu

some website designs

A few years ago I started a web design firm with my brother.

climate-institutions.org

landing and conference page for the Institutions & Effective Climate Action project

chocchique.ca

marketing and e-commerce page for ChocChique, a local bean-to-bar chocolate startup

riftium.ca

3d landing page and contact for Riftium, my local web firm

immanuellaw.ca

landing page and contact for Immanuel Law, a local law firm

selected personal/​​hackathon projects

...from a while back (2019-2023).

🌙 moonlit
emotional news detection

|

3rd @ hack the change

🚫 reBlock🚀 trajectify🎊 reSolve
new years resolutions generation

|

1st @ see you later, hackulator

✈️ triptrove
travel plans generation

|

1st @ wafflehacks


🎮 nimble3d game engine with GUI, physics, and ECS system️️🦀 crabfishchess engine that beats me pretty badly🐱 lulzfast and spec-compliant LOLCODE implementation🐛 issuebasesite to list good first issues in notable github repos💎 zirconinterpreted and purely functional toy language

jonathan li

Reach out to me via email (jonathan@inductionlabs.com).