John Lins

I scraped every startup on TrustMRR, generated high dimensional embedding vectors for each one, then performed a 3D Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction, then normalized into unit vectors centered around a sphere, performed clustering, and finally re-scaled the height of each point based on the MRR of each startup.

Fizzbase is an AI search platform that indexed every person in the Bay Area and mapped how they’re connected.

Banking for clubs or events: collect dues, issue cards, and hand off ownership cleanly without Venmo compliance chaos.