Born and raised in San José, California. Family from Mexico. A student at Stanford with an interest in computer science and how AI can be used to solve problems. I founded NaLi because ~70% of Peru works informally, pays their bills, runs real businesses, and banks can't see any of it. The data is there, and with AI it's a matter of understanding data better. We built it, shipped it, and proved it on a live cohort.
Otherwise: intramural volleyball captain, Ballet Folklórico dancer, founder of Stanford's Migrant Farmworker Drive.
Underwriting infrastructure for informal SMEs · 2025–26
Alt-data underwriting for LATAM lenders. Ingestion connectors normalized Yape/Plin flows, bureau pulls, and device/psychometric signals into a feature store. Linear regression with signed coefficients scored borrowers on a 300–900 scale, served via a decision API: approve/decline, limit, term, reason codes. Evaluated on AUC/Gini against a live Lima cohort. Founder & CEO.
CDAO · AI/ML · Commercial & Investment Bank
As a Forward Deployed Engineering intern within JPMorgan's Chief Data & Analytics Office, I built AI agents for a BlackRock post-trade settlement use case, translating complex operational workflows into agentic systems. I gained hands-on experience with Claude Code, Codex, RAG, knowledge bases, MCP integrations, and enterprise AI tooling. I also worked across AI use case portfolio management and knowledge management, helping teams evaluate, structure, and scale AI use cases.
A bilingual AI caseworker for Latino families
Merlucas uses voice and document intelligence to guide Latino families through taxes, FAFSA, and public-benefit forms — extracting and validating information, flagging missing documentation, tracking deadlines, and escalating complex cases to qualified professionals.
Named for my grandmother in Mexico, who never learned English and called my brother and me merlucas — her own blend of Marina and Lucas. She kept her family afloat through paperwork no one ever explained to her. We built this for people like her.
Turning geopolitical noise into a trade signal
An agent that scrapes geopolitical news sources, models how those events could disrupt global supply chains, and outputs a ranked list of stocks to buy or sell based on the predicted impact.
A few other things I've done, briefly.
Published in San José Spotlight: "Arriola: Cindy Chavez Versus the Concrete Ceiling" (Feb 2023)