Portfolio

Selected Projects

A showcase of real solutions I've designed and built — from cloud architectures to AI-powered tools.

Web & CloudThis site

gaelperez.mx

Consulting platform with built-in video calls (no Zoom or external links), structured per-project onboarding, booking system with email confirmation, and multi-language support. Reusable component architecture built to scale.

Built-in video callsStructured onboardingEmail bookingMulti-language
Next.jsTypeScriptTailwindCSSVideoSDKFastAPI
PlatformiPaaS · LGMDFP compliance

Enlace Escudo — Regulatory Compliance iPaaS

Multi-tenant SaaS iPaaS connecting Mexican institutions to the government PUI platform (RENAPO/SEGOB) for LGMDFP Art. 12 Bis compliance — missing persons search across institutional registries. Privacy-by-design: stores only SHA-256(CURP+RFC) hashes, never raw personal data. 8 Go microservices on ECS Fargate, 3 Next.js 16 frontends, AES-256-GCM + per-tenant KMS CMK, PostgreSQL RLS, full Terraform IaC. Engineering rigor: 124 unit tests under Go's race detector, a 23-phase integration suite attack-testing the OWASP API Security Top 10 (BOLA, SSRF, mass assignment, broken auth, SQLi), per-service OWASP ZAP DAST gated to zero findings, plus gosec/Semgrep SAST and Snyk SCA in CI/CD.

Privacy-by-designPer-tenant KMS CMK8 Go microservicesOWASP Top 10 attack-tested
GoNext.js 16PostgreSQLAWS ECS FargateAWS KMSAmazon SQSTerraformDocker
Nirezumiii — Digital Platform for Art Studio
Web & Cloud100% digital ops

Nirezumiii — Digital Platform for Art Studio

Full digital ecosystem for a tattoo and nail-art studio in Mexico City: WhatsApp bot for appointment management, admin panel, AWS S3 portfolio gallery, automated review system via WhatsApp, and a Next.js 15 frontend with Y2K cyber kawaii aesthetic.

WhatsApp bot + bookingsAdmin panelS3 galleryAuto reviews
Next.js 15Node.jsPostgreSQLTwilioAWS S3RailwayVercel
Mobile & IoT−40% manual entry

Inventory Management System with Specialized Hardware

Mobile and web app integrated with specialized sensors for automatic inventory reading, eliminating 40% of manual data entry in warehouse operations.

Automatic readingMobile + web appHardware integration
Android (Java)Node.jsAWSPostgreSQL
AI & Data~82% accuracy

RAG-based Sentiment Detection

Text sentiment analysis system using Retrieval-Augmented Generation, LLMs and Hugging Face, achieving ~82% accuracy on text classification.

RAG pipelineLLM fine-tuning~82% accuracy
PythonLLMsHugging FaceRAG
AI & Data−40% errors

Visual Computing for Logistics

YOLO-based system for trailer detection and classification in warehouses. Reduced logging errors by 40% and saved 40% in inspection time.

Real-time detectionYOLO modelCloud deployed
FlaskYOLOOpenCVDockerAWS
PlatformFull management

Restaurant Management Platform

Comprehensive system for restaurant operations management, including inventory, orders and reports; integrates AI and economic models for sales cycle prediction and inventory optimization.

Sales forecastingInventory optimizationOperations dashboard
PythonReactFlaskNode.jsDockerAWSNumPy
AI & DataGrowth prediction

Microbiology Culture Management & Growth Prediction System

Data visualization and analysis platform for microbiology culture management, with growth prediction capabilities and environmental condition optimization.

Growth predictionData visualizationEnvironmental optimization
PythonPandasScikit-learnStreamlitSpring Boot (Java)