“FlipArb is a marketplace intelligence system I built from experience repairing and reselling phones. It combines marketplace data, repair pricing, device status checks, AI assisted analysis, and Bayesian learning to identify opportunities worth reviewing quickly.”
“The sourcing view brings listing price, projected profit, shipping, seller information, and listing age into one place so I do not have to manually inspect every result.”
“Each listing passes through cheap filters first. Stronger candidates then receive comparable sale analysis, repair cost estimation, device status checks, AI assisted verification, and scoring.”
“The scanner also learns which search queries are most productive. I used Thompson Sampling so queries that repeatedly surface stronger opportunities receive more scanning budget while weaker queries gradually receive less.”
“Across development and testing, the project scanned more than 200,000 listings. One recorded Deal Engine session contained 120,546 raw listings, 105,606 listings that passed initial filtering, and 26,147 listings that reached analysis.”
“The project was not only a simulation. I purchased devices surfaced by the system, had them shipped to me, and later resold them. That created a real feedback loop between analytical output and resale decisions.”
“FlipArb was developed through the University of Delaware VentureOn entrepreneurship program and combines business analytics with Python, JavaScript, APIs, Bayesian learning, and real resale operations.”