I'm a defense-first security engineer with an offensive backbone — the kind of analyst who reads detection rules from both sides of the wire. My day-to-day spans SOC operations, threat hunting, incident response, penetration testing, and AI/LLM security research.

My experience is anchored in three areas:

• 🛡️ A HomeLab SOC (OPNsense + Suricata + Wazuh + Active Directory) used for hands-on red/blue exercises
• 🔬 FIU cybersecurity research spanning adversarial AI, LLM jailbreak testing, and behavioral ML
• 🐛 Responsible vulnerability disclosure — 150+ findings reported through NCIIPC, HackerOne, Bugcrowd, and OpenBugBounty

I care about measurable security outcomes, not security theater. Every project below has hard numbers behind it.

End-to-end enterprise SOC built from scratch — segmented VLANs (DMZ/LAN/Mgmt), Suricata IDS/IPS, Wazuh SIEM with custom alert rules, Active Directory with GPO + RBAC, and red/blue exercises from an isolated Kali VM.

SOAR-adjacent threat intelligence aggregator — RSS scraping, web scraping, VirusTotal IoC enrichment (hashes, IPs, domains, URLs), MongoDB persistence, scheduled weekly digest delivery. Hardened with Helmet.js, CORS, rate-limiting, RBAC, HTTPS.

🎯 Direct application of SOC analyst IoC workflow at scale.

Novel offensive AI research introducing cryptographic prompt injection as a new jailbreak class against text-to-image diffusion models.

📈 +35% attack success rate · +45% detection efficiency · −60% testing time.

Full PCI DSS SAQ D compliance implementation — Azure Key Vault secrets management, enforced HTTPS, firewall rules for CDE, GitHub Actions CI/CD with secret scanning.

✅ Audit-ready under structured regulatory framework.

50+ web/API assessments mapped to OWASP Top 10 with full kill-chain documentation, CVSS scoring, and remediation guidance.

📤 Methodology behind 150+ responsibly disclosed vulnerabilities.

Digital forensics tooling — EXIF metadata extraction + Isolation Forest anomaly detection to surface tampered or AI-generated imagery.

🔎 Bridges classical forensics with ML-driven anomaly detection.

Dual-environment IAM architecture with dev/prod isolation, least-privilege policies, group-based RBAC, and end-to-end permission-boundary validation.

Hands-on administration across Active Directory, Group Policy, MFA/Conditional Access, Intune MDM, Microsoft 365, Exchange Online, endpoint imaging, patching, and SLA-driven Tier 1/2 support.

🎯 Supported 300+ users and endpoints while sustaining ~92% first-contact resolution.