AI Red Teaming • Adversarial ML • Secure AI Systems • Threat Detection • Multi-Agent MCP Architectures

• 🏆 Best Talk – AI Resilience Hub, Ekoparty 2025
• 🎤 Speaker at DEF CON, Ekoparty, DragonJARCON, PyCon, BSides
• 📚 Author of three cybersecurity & AI books, published by Editorial ICESI
• 🧪 Researcher in adversarial machine learning, AI robustness & red teaming
• 🛠️ Developer of multi-agent architectures using MCP, LangGraph & secure toolchains
• 👨‍🏫 Lecturer at Icesi University, teaching AI & cybersecurity
• 🌐 11+ years delivering end-to-end AI, security & data systems for public and private institutions

I am an AI security–focused data scientist and researcher with over 11 years of experience building secure, data-driven systems.
My work combines cybersecurity, adversarial machine learning, secure AI deployment, narrative/actor investigation, and multi-agent architectures.

I’ve authored three books on cybersecurity and delivered talks in the biggest security and AI conferences in Latin America and beyond—Ekoparty, DEF CON, DragonJARCON, BSides, PyCon, and more.

At Alethea, I develop MCP-driven AI agent systems for detecting and analyzing online narratives, actors, and influence operations at scale.

• Adversarial Machine Learning (image & text attacks)
• AI Red Teaming & LLM robustness evaluation
• Model Control Protocol (MCP) development
• Multi-agent orchestration using LangGraph
• Threat detection & behavior-based analysis
• Secure model pipelines (LLM toolchains, PG/Snowflake endpoints)
• Responsible AI & safe deployment at scale

Machine Learning & Data Science

Cybersecurity Tools

Software Engineering & Cloud

• Cybersecurity: An Approach Based on Data Science
• Cybersecurity: Data Holds the Answer
• Cybersecurity: Data, the Seeds of Chaos

All published by Editorial ICESI.

Adversarial fuzzing toolkit for LLMs and image classifiers.
Includes jailbreak generation, perturbation pipelines, evaluation metrics, and red-teaming workflows.

AI-powered phishing detection system with Streamlit UI, RLHF feedback loop, and model retraining workflow.

Multi-agent orchestration system using MCP, LangGraph, and secure toolchains.

Matrix-themed AI assistant for cybersecurity training, threat analysis workflows & interactive learning.

• Delivered data security solutions across government, education, and enterprise
• Published scientific papers in cybersecurity, healthcare & machine learning
• Speaker & trainer in over 40 workshops and conferences worldwide