About Me

Hey there!

I'm an Associate AI Safety Scientist in UL Solutions' AI Incubation and Innovation Lab, where I help operationalize UL 3115, the industry's first horizontal, product-level AI safety certification framework. I graduated Cum Laude from Drake University in 2026 with dual majors in Computer Science and Artificial Intelligence and minors in Cybersecurity and Information Systems.

At UL, I evaluate AI-based products against UL 3115's safety principles, including robustness and reliability, risk management, fairness and bias, transparency and explainability, and control and oversight. I also conduct AI risk assessments for the AI tools used across UL, helping the company adopt AI responsibly.

On client engagements and pilots, including large-scale LLM evaluations, I help generate synthetic datasets and build the testing materials and ground-truth references behind them. Across all of it, I apply frameworks like the NIST AI Risk Management Framework, ISO/IEC 42001, the EU AI Act, and AI Verify, and I keep close watch on adversarial risks like prompt injection, hallucination, and evasion attacks.

Outside of work, I keep building, from agentic security tools and open-source ML contributions to my homelab.

Thanks for visiting my page!

Education

Drake University

Computer Science, Artificial Intelligence, Cybersecurity & Information Systems

B.S., May 2026 · Cum Laude

Key Coursework:

  • Natural Language Processing
  • Machine Learning
  • Algorithm Analysis
  • Computer Organization & Assembly
  • Database Management
  • Software Engineering
Skills & Achievements

Technical

Python, SQL, ML frameworks

Academics

3x Dean's List, 2x President's List

Awards

Inaugural Outstanding Student of the Year in Cybersecurity, Drake University (2026)

2nd Place, CCSC Central Plains 2026 Undergraduate Poster Competition (VenomX)

Leadership

Drake Sodexo Catering Student Manager (2024–2026)

Hobbies & Interests

Personal Interests

Homelabbing, 3D Printing, F1 & Football

Hobbies

Exploring the Outdoors, Cooking & Gaming

Fun Fact

I've been to the Boundary Waters Canoe Area in Northern Minnesota 7 times, my longest trek was 23 miles!

What I'm Working On

AI Infrastructure

Currently Learning

Building toward production-grade model serving: high-throughput inference with vLLM and container orchestration with Kubernetes and lightweight K3s clusters on my homelab.

vLLM Kubernetes K3s

Nemotron-Nano-9B Heretic

Published HuggingFace Model — 4k+ Downloads

Abliterated version of NVIDIA's Nemotron-Nano-9B-v2, produced using the Heretic tool. A 9B-parameter hybrid Mamba-2/Transformer reasoning model with refusals reduced from 94/100 to 9/100.

HuggingFace ML 9B

Heretic

Open Source Contribution — 19.4k Stars

Top-10 contributor to Heretic, the open-source model abliteration tool. Added Nemotron hybrid model support (Mamba2 SSM architecture), originated from a need to run Nemotron 3 locally for VenomX.

Open Source ML Top 10

Conferences

CCSC: Central Plains 2026

Presenter — 2nd Place out of 24 Posters

Presented VenomX, our AI-powered penetration testing assistant, at the Consortium for Computing Sciences in Colleges Central Plains conference. Competed against 23 other posters and placed 2nd alongside teammates Jordan Martin, Khalid Mohammed, and Nick Guyette.

Presented 2nd Place AI

DUCURS 2026

Presenter — April 16, 2026

Presenting VenomX at Drake University's Celebrating Undergraduate Research and Creative Achievement symposium alongside our capstone team.

Presenter Drake

Iowa Technology Summit 2026

Attendee — Technology Association of Iowa

Attended the annual Iowa Technology Summit. Highlight was the "Securing the Age of AI & Digital Workers" panel covering non-human identities, agentic systems, and AI governance — featuring panelists from John Deere, Wright Services, and Data Innovations.

Attended AI Cybersecurity

Technical Skills

Programming Languages

Advanced

Python

NumPy, Pandas, SKlearn, PyTorch, TensorFlow

Intermediate

Java

JavaFX

Intermediate

Web Development

HTML, CSS, SQL

Basic

MIPS, C, Terminal, R, Powershell

AI & Machine Learning

Advanced

Machine Learning

Classification, Regression, Model Abliteration

Advanced

MLOps & Deployment

Docker containerization, Model deployment(Ollama, VLLM), LoRA fine-tuning (Unsloth & Heretic)

Intermediate

Deep Learning

TensorFlow, PyTorch

Cybersecurity

Intermediate

Security Operations

SIEM (LogRhythm), Security Monitoring, NIST Framework, Alert Tuning, Active Directory

Intermediate

Offensive Security

Nmap, Nessus, Metasploit, hping3, nmap, MITRE ATT&CK Wireshark,Penetration Testing, ISEAGE Competition

Advanced

Compliance & Standards

Advanced: EU RED (18031-1, 18031-2) | Basic: UL2900

Networking

Intermediate

Virtualization

Proxmox, Docker, TrueNAS & ZFS

Intermediate

Network Infrastructure

IDF/MDF Management, Network Segmentation, Ubiquiti UniFi

Intermediate

Network Security

Firewall Management (OPNsense), VLANs

Tools & Technologies

Basic

Infrastructure Tools

JAMF Device Management, Intune Management

Intermediate

Security & Identity Management

Logrhythm, Active Directory

Advanced

Development Tools

VSCode

Cloud & DevOps

Intermediate

Cloud Platforms

Azure

Intermediate

Version Control & Containers

Git, Docker

Intermediate

Operating Systems

Linux (Ubuntu, Kali), MacOS, Windows10 & 11

Professional Skills

Advanced

Technical Support

IT Helpdesk Experience, Debugging, Technical Documentation

Advanced

Problem-Solving

Critical Thinking, Analytical Approach, Quick Thinking

Advanced

Communication

Technical Writing, Presentation Skills, Project Coordination

Additional Skills

Intermediate

Data Visualization

R, Excel

Advanced

Softwares

Office Suite & Google Suite

Projects

VenomX AI-powered penetration testing assistant

VenomX — AI-Powered Penetration Testing Assistant

Multi-specialist agentic penetration testing system with 11 AI agents, a RAG pipeline over 487,327 CVE records from 13 sources, and CVSS-weighted attack path scoring. Runs fully locally — no cloud API required. Presented at CCSC Central Plains 2026, placing 2nd out of 24 posters.

Multi-Agent AI RAG / pgvector vLLM Nemotron 30B Python
VIEW DETAILS
Bank Failure Research project poster showing machine learning architecture for predicting bank failures

Machine Learning for Bank Failure Research: Enhancing Early Warning Systems

In this group research project, we are attempting to explore how machine learning models can best predict bank failures by evaluating financial health indicators. Our goal is to contribute to more reliable early warning systems for banking regulators, enabling proactive intervention that reduces systemic financial risk while providing financial institutions with more accurate, data-driven risk assessments.

Deep Learning Architectures Time Sequence Class Imbalance Handling (SMOTE)
VIEW DETAILS
UL Solutions Cybersecurity Internship presentation showing EU RED compliance documentation and LLM research

UL Solutions Cybersecurity Internship

During my 12-week cybersecurity internship at UL Solutions, I created comprehensive testing documentation for EU Radio Equipment Directive (RED) compliance, specifically developing detailed guides for 18031-1 (Internet-Connected Radio Equipment) and 18031-2 (Personal Data Processing Equipment) testing procedures. Additionally, I researched and developed a proof-of-concept local LLM solution using LoRA fine-tuning techniques to automate documentation generation for cybersecurity testing standards.

18031/RED Certification Local LLM Fine-tuning Pentesting
VIEW DETAILS
Drake University team holding 2nd place trophy at C3 2025 Cyber Defense Competition

C3 2025 Cyber Defense Competition - 2nd Place

I led Drake University's team to 2nd place in the C3 2025 Cyber Defense Competition, designing a three-tier segmented network architecture, creating comprehensive white team documentation, and removing extensive Red Team persistence mechanisms across Windows and Linux systems.

OPNsense Firewall Network Segmentation Active Directory System Hardening
VIEW DETAILS

Recent Activity

GitHub Contributions

@cpagac

Building in public, one commit at a time

Documentation & Guides

Technical Documentation

Guides, tutorials, and notes from my learning journey

Work Experience

Jun 2026 - Present

Associate AI Safety Scientist

UL Solutions · AI Incubation & Innovation Lab

  • UL 3115 AI safety conformity assessments
  • AI risk assessments for AI tools used across UL
  • Synthetic datasets & ground truth for LLM evaluations
  • NIST AI RMF, ISO/IEC 42001, EU AI Act, AI Verify
  • Adversarial risk: prompt injection, hallucination
Mar 2025 - May 2026

SIEM Student Analyst

Drake University

  • SIEM alert tuning & analysis
  • Security monitoring
May - Aug 2025

Cyber Security Engineering Intern

UL Solutions

  • EU RED 18031 documentation
  • LoRA LLM development
Summer 2024

Technology Intern

Glenview School District

  • IDF/MDF infrastructure
  • 1500+ device imaging
Aug 2023 - May 2025
▸ Student Tech
↓
★ Imaging Tech

Drake ITS

  • Tech support & troubleshooting
  • Advanced device imaging
Jan 2023 - May 2026
▸ Student Associate
↓

Drake Sodexo

  • Event catering operations
  • Team leadership & training

Contact Me

Something caught your eye? Send me an email!

Send Message