Predictive X-Ray

Predictive X-Ray

Castro established Predictive X-Ray in 2021, grafting his deep-learning expertise onto Brazil's vastly underserved market for real-time operational intelligence. The firm sits at the intersection of data engineering and domain-specific AI, addressing structural problems in credit underwriting and anti-fraud for a financial system that still relies heavily on manual review and static scoring models. The initial product suite targeted mid-tier Brazilian banks and private insurers, verticals where the cost of false positives in fraud detection and the capital drag of mispriced credit were most acute. Predictive X-Ray's strategy is anchored in enterprise-grade software-as-a-service delivery, deploying models that ingest transactional, behavioral, and third-party data streams to generate anomaly scores and risk flags. While the firm does not publish a client roster, its positioning suggests active engagements within the Brazilian financial services ecosystem, including banks, digital lenders, and health insurers. Deployment is characterized by deep integration rather than broad platform sales — each client engagement typically involves model customization for local data schemas and regulatory classification rules under Brazil's LGPD and Central Bank (Bacen) reporting requirements. The firm competes with both legacy data bureaus and newer Latin American fintech infrastructure players. The firm remains small by headcount, with core technical and go-to-market roles concentrated in Rio de Janeiro and São Paulo. Castro maintains a low public profile, typical of a pre-scale B2B AI company in Latin America, and the firm has not disclosed outside capital raises, suggesting bootstrapped operations or angel backing. In early 2024, the company expanded its public documentation on model interpretability — a functional requirement for regulated Brazilian financial institutions — signaling a maturation of its compliance-ready product layer. There are no known adjacent philanthropic vehicles or real-asset arms. Predictive X-Ray's structural differentiator is its technical embeddedness in Brazil's regulatory-computational gap: the Central Bank's open-finance mandate and Pix instant-payment rails generate vast new data streams, but most incumbents lack the in-house ML capability to act on them in real time. The firm's product essentially functions as an outsourced prediction layer for institutions that cannot build one themselves, a model that becomes more defensible as local data-privacy regulations raise the cost of offshore model training on Brazilian customer data.