Jared Heyman built Rebel Fund's investing approach around a machine learning system that studies millions of data points on past Y Combinator companies, leading him to find that founder background and product progress predict success better than team pedigree alone. His focus on data goes back years before Rebel Fund, from patents on predicting medical diagnoses and product success to founding CrowdMed after his sister's long struggle to get diagnosed with a rare disorder, and today he applies similar data-backed thinking to portfolio building, using math like Monte Carlo modeling to challenge common assumptions about which YC startups are worth backing.