Summary
Garreth Lee is a founder and machine learning engineer focused on building the data layer for perceptual and multilingual AI, currently scaling Mundo AI after joining Y Combinator’s W25 batch. With six years of experience across Cohere, Hugging Face, Wealthsimple and research at UBC, he’s shipped robust pipelines and LLM-pretraining data engineering that bridge research and production. He has hands-on expertise in tokenization and web text extraction, and has applied LLMs to internal data products to boost productivity. Known for turning messy web and multilingual data into reliable training assets, he combines product-minded engineering with research-grade rigor. Based in San Francisco, he pairs startup founder grit with a formal background in computer science and statistics from UBC. An adventurous collaborator who once “swam in the Caribbean waters” while experimenting with tokenizers, he brings both technical depth and unconventional curiosity to ML data problems.
6 years of coding experience
3 years of employment as a software developer
Y Combinator
Bachelor of Science - BS, Computer Science and Statistics, Dean’s List, Science Scholar, Bachelor of Science - BS, Computer Science and Statistics, Dean’s List, Science Scholar at The University of British Columbia
Indonesian, English