Applied AI & reproducible ML
Machine learning · transfer learning · evaluation · traceability · automation
About
Applied AI, GeoAI and technology architecture grounded in real systems.
Martin Heriberto Perez Gomez combines doctoral research, software architecture, data engineering and institutional technology experience to turn ambiguous problems into verifiable, maintainable systems.
Profile
Martín Heriberto Pérez Gómez
Founder & Principal Architect · PhD researcher · Applied AI & GeoAI
Technology architect focused on complex institutional systems that combine structured data capture, validation, storage, analytical processing, geospatial intelligence and decision-support interfaces.
His current doctoral research at CIC-IPN studies whether predictive models can transfer across heterogeneous geographic regions, with emphasis on cross-city transfer learning, distribution shift, negative transfer and reproducible evaluation.
Professional experience spans institutional technology, public-sector data, international cooperation, business intelligence, GIS and software platforms, including work with Pacto Global Mexico, UNODC and INEGI.
Core areas
A single practice across research and engineering
GeoAI & spatial intelligence
GIS · geospatial data · spatio-temporal analysis · cross-region modeling
Technology architecture
APIs · databases · modular systems · integration · operational reliability
Data governance
Structured data · validation · interoperability · provenance
Business intelligence
Decision support · KPI design · dashboards · analytical workflows
Institutional systems
Technology for public, international and organizational environments
Research
Cross-city transfer learning and GeoAI
The research program examines when models trained in one urban context can generalize to another, and when geographic heterogeneity causes transfer failure. Current work connects Mexico City, Chicago and London through reproducible spatio-temporal evaluation.
Evidence
Externally verifiable identity
Academic identity is cross-linked with ORCID, CIC-IPN institutional records, IEEE activity and public research sources rather than relying only on this website.
Open research and evidence →