Cybersecurity & Backend Engineer · Cloud Security · Go · Linux · Open Source

Building secure systems for AI agents and teaching practical technology through VT Security.

I am a Cybersecurity and Backend Engineer based in Rosario, Argentina. I build and secure cloud-backed systems, with a particular interest in identity, permissions, observability, and the Linux and networking layers underneath modern abstractions.

Alongside my professional work, I create open-source software and practical Spanish-language education through VT Security.

I write long-form technical notes at vtsecurity.com.ar - Linux, Git and security, with the diagrams and the exact commands I actually use. I care about work that leaves reviewable evidence: running code, labs, documentation, tests, research, and talks.

Nov. 2025-present

• Work across technical cybersecurity, AWS architecture and security, compliance, and backend engineering.
• Lead a team of four across security and backend work.
• Build Go backend systems and secure AI-agent integrations for controlled access to enterprise data.
• Translate SOC 2 and ISO/IEC 27001 requirements into verifiable technical implementations and evidence.
• Focus on OAuth/OIDC, permissions, sensitive-data boundaries, observability, and operational reliability.

Current focus: AI-agent authorization, MCP security, detection engineering, cloud identity, and low-level Linux fundamentals.

Previously, I worked as a Cybersecurity & Compliance Analyst at Teramot, a Cybersecurity Analyst at Consulting IT, and a Teaching Assistant for Computer Architecture II at Universidad Abierta Interamericana.

• Open Security Labs - My flagship open-source education project: practical Spanish-language labs covering Linux, networking, backend engineering, cloud, DevSecOps, AI agents, and cybersecurity. Explore the platform →
• VT-SecretShare - Self-hosted, one-time secret sharing with browser-side AES-256-GCM encryption, a Go API, and an atomic burn-after-read in Redis (a single Lua check-and-delete script). Try it →
• homelab-blueprint - How I run a single-server homelab like a small production system, as a sanitized, runnable blueprint: Gatus probes and a heartbeat for every scheduled job, a weekly test that restores every backup into a throwaway scratch, a nightly drift check of installed files against git by blob SHA, and a cold standby in AWS with a guarded, rehearsed takeover.
• AutoConfine - Builds least-privilege seccomp profiles for OCI containers from what the container actually does: an eBPF probe records every syscall from the entrypoint's execve on, and AutoConfine turns that trace into a profile that allows only those. Ships as a CLI (go install) and a GitHub Action.
• VT-Agent-Firewall - A local security gateway between an AI agent and its tools (filesystem, shell, network, MCP): default-deny policy engine, human approval, and an append-only audit log, tested against indirect prompt injection. It is a security lab, not a production sandbox: it enforces policy around the tool call, not at the operating-system boundary.
• pluma - The open-source blog platform behind vtsecurity.com.ar and yaninacolombero.com: per-tenant contracts, isolated deployments, content sanitization, and per-instance validation. Tenant selection happens at build time, so each site ships as its own deployment.
• VT-Lens - A native Rust GUI for understanding local process and network activity, turning a selected slice of evidence into an LLM-ready prompt or Markdown export. It reads /proc without root and is an educational instrument, not an EDR or a packet sniffer.
• VT-Ragnaros - A reverse-engineered, from-scratch Linux driver and daemon for the Ragnaros USB control deck: libusb transport, 10 LCD keys, 4 rotary knobs, a touch strip, and YAML profiles. It runs entirely in user space; there is no kernel module.
• VT-Security-Fixes - DEP-3-formatted security patches for open-source packages that miss standard distribution updates, currently focused on Ubuntu universe CVEs.
• VT-Terminal-Project - A reproducible macOS and Linux terminal environment with Zsh, Ghostty, modern CLI tools, searchable cheatsheets, and AI-assisted workflows. Its ide/ folder holds an AI-native Zed setup with MCP integrations, inline assistance, and structured Git review, conflict-resolution, and diff workflows. Versioned releases since v1.0.0, with a one-line install.

My merged Podman contribution, Add --retry support to manifest push - PR #28637, added --retry and --retry-delay across the CLI, local and remote clients, REST API, manpages, Swagger documentation, and end-to-end tests.

It is a small feature with a production lesson behind it: reliable infrastructure requires explicit behavior for temporary network and registry failures.

These are the technologies I use across secure backend systems, cloud infrastructure, identity, observability, and systems engineering.

Backend and data

Cloud and operations

Security and compliance

Linux, containers, and networking

• Information Systems Engineering, Universidad Abierta Interamericana - final year, expected Dec. 2026
• Information Systems Analyst, Universidad Abierta Interamericana - completed Dec. 2024, GPA 9.25/10
• Teaching module for university teaching assistants, Universidad Abierta Interamericana - module certificate, 2026
• CompTIA Security+ (SY0-701) - issued Feb. 2026, valid through Feb. 2029
• AWS Certified Cloud Practitioner (CLF-C02) - issued Jul. 2025, valid through Jul. 2028
• Linux Foundation Certified IT Associate (LFCA) - issued Apr. 2026, valid through Apr. 2028
• Huawei Certified ICT Associate, Datacom (HCIA-Datacom)

VT Security is my Spanish-language education and community platform for practical cybersecurity, Linux, networking, software engineering, and systems fundamentals. Open Security Labs is its flagship public software project.

I am writing my first book, No podés defender lo que no entendés (in Spanish): a path into cybersecurity that starts by understanding Linux, from the command line down to the kernel and up through containers, the cloud and AI agents, with Open Security Labs labs for its chapters. Join the waitlist →

I am an AWS Community Builder, in my first year in the programme.

• JCC XXIV 2026, FCEIA, UNR, Oct. 21, 2026: invited 45-minute talk, El problema no es el prompt: seguridad real en aplicaciones con IA

• Ekoparty 2026, Buenos Aires, Oct. 7-9, 2026: five talks: Prompt injection en agentes con tools (AI Resilience Hub), Dónde se rompe OAuth cuando el que llama es un agente (OWASP Village), Cuando el webhook llega dos veces (CyberFinance), Logs bajo fuego (BlueSpace) and GitOps: cuando la fuente de verdad también es el riesgo (DevSecOps Space)
• CACIC 2026, UTN FRCU, Oct. 6, 2026: presented two accepted papers virtually, on Podman/FOSS contribution and on Ubuntu universe SRU work
• Hacking Day 2026, Paraná, Oct. 2, 2026: Firewall para agentes de IA: prompt injection en vivo (main stage, 45 min)
• Joven Argentina (FNGA), Rosario, Sep. 28, 2026: cybersecurity talk and exchange for a programme of young leaders
• UAI Rosario, guest lecture, Sep. 25, 2026: Sistemas operativos Linux
• Webinar UAI, Tecnología Informática, Aug. 19, 2026: Orquestación, workers y arquitectura moderna: cómo diseñar sistemas que escalan
• DebConf26, July 24, 2026: Abstraction Leaks: Why Understanding Linux Internals Still Matters
• Vincular Inteligente 2026, May 22, 2026: Seguridad con IA: monitoreo inteligente y respuesta temprana
• CyberSecTuc Meetup #3, May 2026: La realidad de un Ingeniero en Ciberseguridad
• CACIC 2024: Expositor Distinguido en Seguridad Informática for reverse-shell research
• SACS / 53 JAIIO 2024: Mejor Exposición for botnet taxonomy research
• JAIIO 2024: co-author of research on monoliths and microservices
• WICC 2025: Dockerización de servidores SCADA: ciberseguridad industrial
• WICC 2024: Seguridad en redes wifi: estrategias de detección y expulsión de intrusos

Rosario, Argentina · UTC-3