Ernesto Vizcaíno Alvarado
AI / Product Engineer
I build AI-powered products that solve real-world problems.
I've led engineering for fintech startups, shipped software used by thousands, and developed machine learning systems for astronomical research. My work spans full-stack engineering, applied AI, and product development, from early prototypes to production systems.
Pressto copy my emailCopied
Selected work
Selected products, systems and research I've built.
- Applied AI · FintechNext.js · Vercel AI SDK · AWS · BigQueryLed development of a production fintech application for loan origination, risk assessment, compliance, and financial document processing. Built agents with the Vercel AI SDK, integrated financial data, and developed AWS services and BigQuery pipelines for customers including Abaco Capital.
Oliver AI — Financial AI Agents
- (opens in a new tab)Applied AI · Data EngineeringMXN $4M+ in SMB financing in the first monthBuilt credit preapproval, AI agents, and knowledge bases using financial, banking, and tax data. Developed CONTPAQi integrations and accounting ETL pipelines for automated reconciliation and Excel reporting.
Xignus AI — Financial Automation
- (opens in a new tab)Machine Learning · Astronomy5.1M+ objects · Time-series features · Model evaluationBuilt an uncertainty-aware pipeline for time-series classification, anomaly detection, and scientific prioritization across 5.1M+ astronomical objects. Used feature engineering and evaluation on unseen observing cadences to assess morphology classification without physical ground-truth labels.
Astronomical Machine Learning — Rubin/LSST
- Product Engineering · SaaS10,000+ downloads · 3,000 active users in 3 monthsBuilt and shipped a mobile point-of-sale application for small businesses, then expanded it into a web-based ERP. Owned product development from the initial mobile release through the broader business-management platform.
Oliver POS / ERP
Professional experience
I work across engineering and product, taking ideas from prototype to real users and production.
Independent Product Engineer / Founder
Building and shipping independent software products, from architecture and LLM workflows to backend infrastructure and production deployment. I own API integrations, authentication, payments, SEO, analytics, conversion optimization, and ongoing product iteration and maintenance.
AI Engineer atXignus AI
Built credit preapproval systems that helped SMBs access MXN $4M+ in financing in the first month, plus AI agents and knowledge bases integrating financial, banking, and tax data. Rejoined in August 2026 to build CONTPAQi integrations, accounting ETL pipelines, automated reconciliation, and Excel reporting.
Technical Co-Founder atOliver AI
Led end-to-end engineering of a production fintech application in Next.js. Built AI agents with the Vercel AI SDK for loan origination, risk assessment, compliance, and financial document processing, with AWS services and BigQuery data pipelines serving customers including Abaco Capital.
Founder / Engineer atOliver POS / ERP
Built and launched a mobile POS that reached 10,000+ downloads and 3,000 active users within three months, then expanded it into a web-based ERP.
Research & Awards
1st Place, Xólotl Hackathon
Phase 3, Advanced Track · CUDI / LAMOD UNAM
Rubin/LSST Astronomical Time Series Classification
Won Phase 3 (Advanced Track) with a classification, anomaly-detection, and scientific-prioritization pipeline for Rubin/LSST-like light curves. Evaluated morphology classification on 6,000 light curves from 600 unseen observing cadences; these scores measure morphology classification, not physical ground-truth labels.
- 5.1M+ astronomical objects
- 10 morphologies
- 6,000 evaluation light curves
- 600 completely unseen observing cadences
- 68.47% balanced accuracy
- 68.28% macro F1
View university recognitionThe award supported participation in CARLA 2026 in Córdoba, Argentina.
Research
Gaia / OGLE Variable Star Classification
Universidad Autónoma de San Luis Potosí
Built a machine learning pipeline for periodic variable star classification, processing 491,073 Gaia sources and engineering 91 time series, Fourier, color and catalog features. A class weighted XGBoost model reached 0.9495 macro F1 on held out data and 0.9847 weighted F1 against mapped OGLE labels.
Press & Recognition
University recognition and selected coverage of my work in astronomical machine learning.
Writing
Contact
I'm open to founding, product, and applied AI engineering roles with ambitious teams. Get in touch to discuss what you're building.