I like research that stays close to the real world: teaching tools, accessibility, and how AI is changing the way we work. Each paper below leads with a plain-English summary; expand for the formal abstract.
Publications
We looked at whether the Raspberry Pi 5 can make engineering and CS education more hands-on. A custom Pi-5 kit gave students an affordable, real platform for learning embedded systems, networking, and operating systems, and the data showed it deepened engagement, problem-solving, and resilience, with mentorship and community turning out to matter a lot.
Explores integrating the Raspberry Pi 5 into engineering/CS education for system-level learning. A custom Pi-5 kit provided a cost-effective, hands-on platform for embedded systems, networking, and operating systems. Combining qualitative reflections and quantitative survey data, findings show the Pi fosters deeper engagement, problem-solving skills, and resilience, with mentorship and community support crucial to student success.
We studied how tools like ChatGPT changed conversations on Stack Overflow from 2022–2024. LLM-generated answers are a small but growing share, and they tend to get fewer upvotes than human ones. Humans still write the highest-quality answers, which raises real questions about trust, authenticity, and how forums should moderate AI content.
Analyzes how LLM tools influenced engagement and content quality on Stack Overflow from 2022–2024, combining sentiment analysis, AI-detection methods, and multivariate regression on 3,000 sampled comments. Key finding: a modest but growing presence of LLM-generated responses, which tend to receive fewer upvotes than human-written ones; human contributions still dominate high-quality responses. Highlights the need for better AI-detection tools and raises questions about authenticity, trust, and moderation.
Alongside the engineering, my ScribeAR work was recognized at the UIUC Undergraduate Research Symposium two years running: Outstanding Exhibit (2025) and Honorable Mention (2024). The research covered speaker identification, diarization, and audio classification in service of real-time, accurate captions for accessibility.