Profile banner
Sebastián A. Cruz Romero

Sebastián A. Cruz Romero

CEO and Bioengineering graduate student building Edge AI for health

scruzromero
Mayagüez, Puerto Rico
51 connections
Joined March 2026
Favicon imagecapicu.ai
Favicon imagecemi.capicu.ai
Favicon imagescholar.google.com
Favicon imageyoutube.com
Favicon imageromerocruzsa.github.io

Summary

Builds edge-first AI/ML systems for biomedical settings: Sebastian focuses on making AI models and data systems that run where biological and clinical data are collected, prioritizing offline operation, low-power hardware, model compression, and privacy-preserving local storage so diagnostics and EHR capabilities work in resource-limited environments. capicu+2
Cross-disciplinary researcher with practical deployments: His academic and technical work spans computer science, signal processing, computer vision, and computational biology, and includes peer-reviewed and preprint publications on anemia screening and model quantization targeted for mobile and embedded healthcare applications. google+1
Entrepreneurial leadership translating research to products: As co‑founder and CEO of Capicú Technologies, he leads a team developing tooling (CEMI) and deployed systems (portable EHR/anemia screening) that bridge academic research and real-world deployments, and his venture was selected for VentureWell's Winter 2026 E‑Team cohort. capicu+1
Committed to mentorship and community impact in Puerto Rico: Throughout his undergraduate career he has emphasized mentorship, undergraduate capacity-building (EMB UPRM leadership), and efforts to make rigorous research and tools accessible to underserved settings in Puerto Rico. shinstitute+1

Work

Education

Projects

Writing

Design of an Edge-based Portable EHR System for Anemia Screening in Remote Health Applications

January 1, 2025

Paper describing a portable, edge-enabled EHR platform optimized for offline operation with an integrated non-invasive anemia screening module (fingernail pallor analysis), model quantization for edge deployment, and privacy-preserving local storage.

Favicon imagearxiv.org

Performance Analysis of Post-Training Quantization for CNN-Based Conjunctival Pallor Anemia Detection

January 1, 2025

Study evaluating MobileNet-based conjunctival pallor models on the CP-AnemiC dataset and the effects of post-training quantization (FP32, FP16, INT8, INT4) on diagnostic performance and edge deployment trade-offs.

Favicon imagearxiv.org

Synthesis, characterization, and photocatalytic activity of ZnS and Mn-doped ZnS nanostructures

January 1, 2021

Research article co-authored in 2021 on synthesis and photocatalytic characterization of ZnS and Mn-doped ZnS nanostructures.

Favicon imagescholar.google.com

Hobbies

Mentoring and capacity-building for students and early-career researchers in Puerto Rico. shinstitute