Patrick Memusi

• Relevant coursework: Data Structures & Algorithms, Design Thinking & Communication, Discrete Mathematics, Artificial Intelligence
• GPA: 3.73 / 4.00

AI Open Source Capstone

• Developed and trained a natural language processing model to summarize technical articles, achieving an 85% ROUGE score.
• Engineered a data pipeline to process and clean a dataset of over 10,000 articles for model training.
• Utilized Languages/Tools: Python, PyTorch, Git, Docker

How2 with AI - YouTube Channel

• Created and launched an educational YouTube channel focused on teaching effective AI usage, producing 15+ tutorial videos covering prompt engineering, AI tool workflows, and practical applications.
• Developed comprehensive content strategy targeting beginners to intermediate users, covering topics from ChatGPT optimization to advanced AI automation techniques.
• Grew channel to 500+ subscribers and 10K+ total views within first 3 months, with average engagement rate of 8.5%.
• Utilized Tools: Video editing software, AI platforms (ChatGPT, Claude, Midjourney), analytics tools

How2 Mobile Application

• Designed and built a full-stack mobile app that uses a generative AI model to create educational 'how-to' video scripts from user prompts.
• Implemented a clean, intuitive front-end allowing users to easily generate, edit, and share content.
• Utilized Languages/Tools: JavaScript, React Native, Firebase, Generative AI APIs

Cybersecurity Assessment Report

• Performed a comprehensive security audit on a simulated corporate network, identifying and documenting critical vulnerabilities.
• Conducted penetration testing exercises, including network scanning and vulnerability exploitation, to assess system resilience.
• Utilized Tools: Nmap, Wireshark, Metasploit

• Managed event setups, provided technical support, and collaborated with a team to maintain facilities for 20+ student organizations weekly.

• Investigated how large language models encode and retrieve information compared with human episodic and semantic memory, designing experiments that paired human recall tasks with model prompts.
• Built evaluation pipelines in Python to measure consolidation, interference, and priming effects across 1,500+ trials, visualizing results with reproducible notebooks.
• Analyzed memory decay patterns across different AI architectures & Co-authored a paper summarizing methods, datasets, and ethical considerations around machine memory augmentation

• Organized 30+ meetings and 10+ community events, and led a recruitment drive that onboarded 100+ new members.

• Led a 21-member team through practices and conditioning, resulting in a championship victory in the 2020 Regionals League Cup.