Computer Science & Cybersecurity · Kean University '27
AI security · LLM red-teaming
Undergraduate researcher with Dr. Boyang Li
I'm interested in how AI systems fail, how we test their defenses,
and what evidence we need to trust their outputs.
Seeking Summer 2027 internships
AI security · Security engineering · Applied AI research
- B.S. Computer Science, Cybersecurity Option, Kean University. Graduating May 2027.
- Undergraduate researcher in Boyang Li's group (Team → Current Students), studying verification in decentralized AI systems.
- AI security: directing AI-assisted local experiments on a copied HadAgent prototype and studying methods for verifying LLM inference.
- Peer Mentor / Code Samurai: helping Kean students work through Python, SQL, data structures, and debugging.
- KEAN CTF 2027: co-building a college capture-the-flag competition with teammates and AI assistance.
Tests score validation and model-verification boundaries in a copied research prototype. Includes protocols, result records, and reproducibility instructions.
Python · Local verification tests
Team research comparing prompted and fine-tuned debate coaches. Helped with fine-tuning, evaluation, citation checks, participant interviews, and presentation.
JavaScript · Node.js · OpenAI API
Archived pipeline for prompt-injection, jailbreak, and data-leak tests, with model-judge verdicts and saved transcripts.
Python · Ollama · SQLite
Prepares and validates synthetic tutoring dialogues for fine-tuning, with a three-profile qualitative probe.
JavaScript · Node.js · OpenAI API
Research context & contributions
- HadAgent: I direct research questions and interpret results. AI agents contributed substantially to experiment design, implementation, execution, and review. Findings concern the tested local snapshot, not a deployed network or the complete protocol.
- Adaptive learning: the project uses synthetic dialogues and learner profiles. It does not establish improved student learning outcomes.
Research presentations: NSF IUSE Scholars Academy (2025 and 2026). Presented adaptive-learning response comparisons as first author and a Gen Z housing-demographics study as sole student author at Great Minds in STEM. The studies concern model responses and county-level associations, respectively; neither establishes a causal effect.
| Area | Tools & methods |
|---|---|
| AI evaluation | Ollama · OpenAI API · LLM-as-judge workflows · Citation checking |
| Security | OSINT · OWASP LLM Top 10 · MITRE ATLAS · Adversarial testing |
| Code & data | Python · Java · JavaScript · SQL · PHP · PowerShell |
| Development | Linux · Docker · Git · Node.js · MySQL · SQLite · Arduino/ESP8266 |
| AI-assisted workflow | Claude Code · Codex |
| Coursework exposure | TypeScript · Wireshark · Metasploit · Hashcat · Autopsy |
Selected for Kean University's Work-Earn-Live-Learn (WELL) Program, Summer 2026. WELL combines paid campus work, residential learning, career development, and international study to support student success and timely graduation.
- Cybersecurity internship: investigated nonprofit cyber incidents through OSINT for the Non-Profit Cyber Incident Repository, built a browser-controlled Arduino MEGA/ESP8266 smart-home lab, and taught at the Port Authority STEaM Camp serving 60 high school students.
- Career development: participated in Career Management coursework, resume and interview preparation, financial-literacy workshops, employer panels, and networking with Kean alumni.
- Young Envoys Scholarship Fellow: participated in the June 13-22, 2026 educational and cultural exchange in Wenzhou and Beijing, China, through the China Education Association for International Exchange (CEAIE).
- International speaking: addressed international delegates at a CEAIE event in Beijing and reflected on the experience in my final WELL presentation, connecting technical work with cross-cultural communication and professional growth.
- Great Minds in STEM (2025 & 2026): presented NSF IUSE research on adaptive learning with LLMs (first author) and Gen Z housing demographics (sole student author).
- CRA UR2PhD Undergraduate Mentoring Workshop & Research Showcase (April 2026, New Orleans): presented our team's Debate Coach research and participated in research-career and graduate-school discussions.
- PyCon US (2024): participated in open-source work on Python libraries and developer documentation; selected contributions are linked below.
- IFWY North American Regional Dialogue (September 2-3, 2025, Cornell Tech, Roosevelt Island, NYC): selected as an IFWY Catalyst for the fully funded regional dialogue on environmental policy and sustainable solutions.
- International Forum We, the Youth Final Conference (2025, Seoul): participated in the final conference as one of 25 North American delegates, contributing to environmental-policy and sustainability discussions.
- pyOpenSci: proposed Windows-specific pipx installation instructions for the Hatch tutorial. Recognized as a contributor for code and review.
- PyScript documentation: submitted a correction to the getting-started guide clarifying JSON configuration setup.
- annotated-types: collaborated with cksleigen on a proposed improvement to
GroupedMetadatahandling during my first open-source contributions.
- Kean Center for Cybersecurity: investigated nonprofit cyber incidents through OSINT and taught cryptography, programming, and IP addressing at STEaM Camp.
- WiCyS secretary / PHD4FREE!: coordinated guest-speaker scheduling and event planning from August through October 2026; presented pathways into research, paid undergraduate programs, and graduate fellowships alongside Notre Dame PhD student Demetrius Hernandez for 15 attendees.
- IFWY fellow: received a fully funded fellowship at Cornell Tech, then was voted one of 25 North American delegates for the final conference in Seoul. Led the design and scope of SelectSolar, a proposal to reuse solar panels for educational and nonprofit institutions.
Happy to talk AI security, research, or teaching.
Get in touch · Visit my card