About Me
The Autonomous Defender
I am a Cybersecurity Management Master’s candidate specializing in the intersection of Security Automation, Systems Integration, and Agentic AI. My work focuses on bridging the gap between the cognitive power of generative AI and the strict, deterministic safety required in enterprise security.
My core expertise lies in architecting hybrid-inference pipelines on high-performance infrastructure like the NVIDIA DGX. By orchestrating localized edge models (Llama 3.2) alongside cloud intelligence (GPT-4o), I engineer end-to-end autonomous SOC loops that detect, analyze, and mitigate threats without human intervention—safeguarded by strict programmatic governance to prevent AI-driven outages.
I am currently seeking opportunities in Security Automation, Detection Engineering, or Systems Integration where I can apply autonomous, data-driven strategies to defend critical infrastructure.
Technical Arsenal
| Domain | Skills & Tools |
|---|---|
| Agentic AI & Orchestration | CrewAI, LangChain, Hybrid-LLM Inference, Prompt Engineering, Llama 3.2, OpenAI GPT-4o |
| Security Engineering | Autonomous Incident Response, Active Defense, SIEM, Forensic Triage, Linux Kernel Networking (iptables) |
| Systems Integration | Python (FastAPI/Uvicorn), Docker Containerization, Regex Telemetry Pipelines, RESTful APIs |
| Infrastructure | NVIDIA DGX (ARM64), Azure (AZ-500 Candidate), Active Directory, PowerShell |
🟢 Featured Project: Hybrid-LLM Autonomous SOC & Active Defense Framework
Status: Architecture Complete & Validated
Traditional SOCs suffer from alert fatigue, while unconstrained “AI Security Agents” present massive operational risks to host infrastructure. I engineered a modern solution: an autonomous security loop that handles the entire incident lifecycle while enforcing strict deterministic governance over AI actions.
Technical Milestones:
- Hybrid-LLM Orchestration: Architected a multi-agent pipeline utilizing CrewAI. A local ‘Senior SOC Analyst’ (Llama 3.2) performs privacy-preserving behavioral triage on raw logs, while a ‘Security Auditor’ (GPT-4o) executes external CVE validation and policy compliance.
- Ephemeral Telemetry Pipelines: Integrated isolated Docker honeypots with a custom FastAPI backend, engineering regex-driven Python scripts to dynamically stream attacker IPs from ephemeral containers directly into the cognitive pipeline.
- Deterministic Active Defense: Solved the risk of AI-driven Self-Denial-of-Service (DoS) by engineering a hardcoded Python governance layer. This layer intercepts AI actuation commands and evaluates them against critical infrastructure whitelists before execution.
- Autonomous Actuation: Achieved zero-intervention threat containment by programmatically triggering dynamic
iptablesdrop rules against validated, non-whitelisted malicious endpoints.
Quantifiable Impact:
- Architectural Safety: Proved that AI can be safely deployed for active network defense by separating cognitive reasoning from programmatic execution.
- Instant MTTR: Reduced Mean Time to Respond (MTTR) to near-zero through automated, dynamically governed network isolation.
- Optimized Systems Integration: Successfully unified containerized infrastructure, real-time API streaming, and hybrid AI inference into a single, stable deployment on NVIDIA DGX hardware.