"My industry background shapes how I approach problems, but the questions I care about most live in research."
I started with a B.S. in Computer Sciences at UW-Madison, where I first got into game development and software engineering. During and after my M.S. in CS Engineering at Santa Clara University, I co-founded two startups and spent several years building AI-powered products, from generative AI applications and social platforms to multi-agent systems and intelligent marketing system. We invested heavily in analytics, A/B testing, behavioral data pipelines, and even generative AI systems interpreting our user data, more than most startups at our stage. But even with all that infrastructure, the tools hit a ceiling. We could see what users did, but not the shape of how they were thinking or where they were getting stuck. That gap is what pulled me toward research.
I came to UC Santa Cruz for my PhD in Computational Media at the GUII Lab, advised by Professor Magy Seif El-Nasr. Now I study how AI can understand and support human behavior in games and learning environments. My industry background still shapes how I approach problems, but the questions I care about most live in research.
I've always cared about having real time to think deeply, my family put a lot of weight on that growing up, not just picking up skills but actually reflecting and absorbing. After some time working in industry through internships, I found that building products was rewarding, but I kept encountering questions engineering couldn't answer: why do some players get stuck and just give up? Why does a learning experience click for one person and not another? A PhD felt like the right place to slow down and take those questions seriously, instead of always shipping the next thing.
"I want to create experiences that learn about their players and become better because of it."
When people learn through games and puzzles, what really matters isn't whether they win, it's how they struggle. Most analytics only capture outcomes like scores or completion rates, so the struggle itself gets lost. I focus on the moments where someone clearly isn't getting anywhere but their behavior doesn't change. Maybe they keep repeating a failed approach, or they ditch a strategy that was actually working after one bad result. I'm building a tool that can read those patterns from how people play, so a researcher can actually see where someone got stuck and why.
I've been making games since primary school, I started with board games, writing rules and systems for my friends to play. Later I got hooked on real-time and turn-based strategy games, and what pulled me in wasn't really the mechanics, it was the stories and how personal a game could feel. I kept wondering: what if a game could actually learn about you while you played, and adapt its challenge and pacing to who you are? That's what led me to player modeling, dynamic difficulty, and personalization. When I came across Computational Media and HCI, it just clicked, game design, AI, and understanding people all in one place, where building things and understanding people are the same job.
Honestly, the moment that stuck with me came from a study that went "wrong." We built PEARL, a help agent for our puzzle game that teaches parallel programming. It read the board, pulled the right hint, and offered it at exactly the right time. We assumed players would take good, well-timed help. About half of them just didn't, they'd close the agent and keep going on their own. My first reaction was that the tool had failed. But the longer we sat with the data, the more I realized that was the finding. Whether someone wants help is deeply personal: some people want to solve it themselves, some don't want their flow interrupted. That reframed how I think about my whole research.
"I want to stay at the intersection of games, AI, and learning, whether that means founding a startup, joining a research-driven studio, or continuing in academia."
I want to stay at the intersection of games, AI, and learning, whether that means founding a startup, joining a research-driven studio, or continuing in academia. On the entrepreneurial side, I have co-founded two startups already and I am drawn to building products that make adaptive, personalized experiences real for players and learners. On the research side, I want to keep pushing on questions about how AI can understand and support human behavior in interactive environments. Ultimately, I see myself building things that sit at the boundary of research and product: tools and games that are grounded in serious research but actually reach people. I want to create experiences that learn about their players and become better because of it.
Come in and explore. There are tons of projects you can jump into, and it's a great place to start your own. What makes GUII special is the people, we've got visiting scholars from all over and collaborators across disciplines, so every lab meeting pushes your thinking somewhere new. And Magy is the best advisor I could ask for. She genuinely cares about you as a person, your research output, and she makes it feel safe to take risks. Bring your curiosity and a willingness to collaborate.
This paper proposes that humans understand other people's actions by inferring their goals through Bayesian inverse. The idea that understanding behavior is fundamentally about goal inference, shaped how I think about player modeling. It sits exactly at the intersection of cognitive science and AI that my work lives in.