AI Security Research
Testing emerging LLMs for injection risks, data leakage, and unsafe model behavior.
Hi, my name is Divy Patel. I focus on AI security research, penetration testing, security operations, and automation. I build hands-on projects, test systems for weaknesses, and turn technical work into clear, useful outcomes.
Contact MeName: Divy Patel
Location: Buffalo Grove / Chicago, Illinois
Focus: Cybersecurity, AI security, scripting
Background: Security operations, red team, AD, automation
I am a cybersecurity and risk professional with 3+ years of experience supporting security operations, access control, documentation standards, and technical risk analysis in complex environments.
My work includes AI security research, Active Directory administration, ethical hacking on open-source LLMs, compliance-aligned security practices, and automation for data extraction and analysis.
I also like building personal projects, experimenting with systems, and staying sharp through hands-on learning, especially where cybersecurity, AI, and infrastructure overlap.
Homemade projects and technical work that show how I approach security, analysis, and tooling.
Testing emerging LLMs for injection risks, data leakage, and unsafe model behavior.
Secure red-team tooling built for restricted environments with statistical performance analysis.
Python and Excel workflows for extraction, cleaning, and reporting.
Collision simulation with parameter sweeps, modeling, and visualization.
Practical help for security, troubleshooting, and technical improvement.
Security reviews, risk analysis, and guidance for technical environments.
Assessment of LLM workflows, policy gaps, and model risk considerations.
Python-based extraction, cleaning, validation, and reporting automation.
Device troubleshooting, setup, and general technical support for local clients.
National Institute for Aviation Research
Lead AI security research, test LLM vulnerabilities, manage Active Directory access, and write internal AI security protocols.
Built a secure offline system, analyzed password-cracking performance, and developed data pipelines for operational research.
Directed budget planning, outreach, chapter communication, and cybersecurity-focused workshop coordination.
National Institute for Aviation Research — automated extraction and cleaning, supported DoD research, and improved cataloging workflows.
Lets connect about cybersecurity, AI security, automation, or repair support.