Awardees

Safe AI-Enabled Real-Time Control of Hydraulic Fracturing via Neural Operator Surrogates and Spectral Safety Projection

My research focuses on building neural operator surrogate models for hydraulic fracturing and using them inside a differentiable predictive control framework to find the best pumping schedules in real time. A key part of the work is a spectral safety projection layer that gives the controller formal guarantees, so it can run much faster than traditional methods without trading off safety. The same approach can be extended to other PDE-governed civil engineering problems, such as dam seepage, ground deformation, and underground energy storage.

Why did you become a scientist?

I have always been curious about how things work and why they happen the way they do. Even as a kid, I would take things apart just to see what was inside. That same curiosity, applied to bigger questions, is what pushed me toward research. Becoming a scientist felt like the most natural way to keep asking those questions and actually try to answer them.

What is your favorite aspect of your research?

My favorite part is the moment when I dig deep into a concept and notice something small that other people missed. It is a quiet kind of excitement, but it is what I look forward to the most. Often, that one detail opens up a new way of looking at the problem, and the work starts to feel like mine.

What excites you the most about the opportunity to work on AI and Civil Engineering?

AI is moving very fast, but most of the impressive results so far live in the digital world. Civil engineering is the opposite. The systems are physical, slow to build, and very hard to fix once something goes wrong. Bringing AI into this space is exciting because the stakes are real. A neural operator that runs in milliseconds can replace simulations that take hours, and a well-designed controller can help prevent serious issues like aquifer contamination or wellbore failure. I find it meaningful to work on something where better AI translates directly into safer infrastructure.

What do you hope to accomplish during your time as a Chishiki AI Fellow?

I hope to build a complete, working version of my framework, from the neural operator surrogate to the safe controller, and release it openly so other researchers can use it. The HPC support from Chishiki AI matters a lot for this, since the GEOS simulations alone need a large amount of compute. I also want to use this time to learn from other fellows and faculty, share my work with people outside my immediate field, and publish in strong venues. If by the end of the fellowship, the methodology can be picked up and applied to other civil engineering problems like dam seepage or underground energy storage, I would consider it a real success.

What do you like to do when you aren’t working on research?

When I am not working on research, I like spending time with my friends and trying out new recipes in the kitchen. Cooking is my way of switching off. It is hands-on, the feedback is immediate, and there is no peer review.

Trustworthy Adaptive Experimentation for Civil-Engineering Digital Twins

My research develops trustworthy AI methods for civil-engineering systems that must learn and make decisions under uncertainty. I focus on the problem of how engineering systems should decide what to simulate, measure, or reconstruct next when prediction, sensing, and model updating are tightly coupled but often treated separately. To address this, I combine Bayesian optimization, uncertainty-aware scientific foundation models, and digital-twin thinking to build adaptive systems that are not only informative, but reliable and useful for real engineering decisions.

Why did you become a scientist?

I became a scientist because I realized early that research and innovation are the foundation of resilient and prosperous societies. My path was shaped both by my civil-engineering background and by personal experience, including watching my father work as a foreman on construction sites and later seeing the realities of engineering practice for myself. Over time, what stayed with me most was the decision-making problem: how do we make better choices in complex, high-stakes settings where uncertainty is unavoidable? That is still what drives me. I want to solve problems that move from research into real value for engineering systems and the communities they serve.

What is your favorite aspect of your research?

My favorite aspect of research is helping bring the most important technological advances of our time into civil engineering in a principled way. Civil engineering operates in complex and high-risk settings, but it has not always benefited quickly from new technologies. I enjoy working on methods that make modern AI and uncertainty quantification truly compatible with these settings. I especially enjoy releasing open-source tools, presenting my work at leading machine-learning venues, and teaching others about it, because those are the moments when research begins to move beyond papers and into broader impact.

What excites you the most about the opportunity to work on AI and Civil Engineering?

What excites me most is that AI opens the door to a more complete vision of digital twins in civil engineering systems that combine learning, fidelity, intelligence, and closed-loop decision-making. As the world begins to pay more attention to physical AI, I believe this is an especially important moment for civil engineering. My work has increasingly focused on adaptive and optimal experimental design, and I am excited by the possibility of using AI not just to model engineering systems, but to help them learn, adapt, and make better decisions in technically meaningful ways.

What do you hope to accomplish during your time as a Chishiki AI Fellow?

During my time as a Chishiki AI Fellow, I hope to accelerate my research on trustworthy adaptive experimentation for civil-engineering digital twins. I want to deepen this work, expand its impact, and move it closer to the kind of reliable, closed-loop intelligence that can make a real difference in engineering practice. I am also excited about the opportunity to learn from mentors, collaborate with other fellows, and fully leverage advanced computing resources to pursue this research in a more ambitious and unhindered way.

What do you like to do when you aren’t working on research?

When I am not working on research, I enjoy church service and giving back to my community. I also like exploring the venture capital space, listening to diplomatic speeches, and watching a good football (soccer) game.

AI-SLIDE — An Artificial Intelligence Assisted Spatiotemporal Landslide and Debris-flow Risk Assessment Tool

My research develops AI-SLIDE, an artificial intelligence and cyberinfrastructure-enabled framework for modeling landslides and post-fire debris flows across space and time. The framework integrates large-scale geospatial data, synthetic rainfall generation, fragility modeling, and transportation network analysis to identify vulnerable road segments, estimate mobility disruptions, and support equitable infrastructure investment decisions. While my current focus is on landslides and debris flows, my long-term goal is to expand this framework into a broader AI-assisted decision-support platform for multiple natural hazards affecting civil infrastructure and community resilience.

Why did you become a scientist?

I became a scientist because I have always been motivated by the combination of analytical problem solving and community impact. Civil engineering research allows me to work on technically complex problems while also addressing real risks that affect people’s daily lives, mobility, safety, and access to essential services. I am also deeply passionate about asking research questions and developing tools that can help communities become more resilient to natural hazards.

What is your favorite aspect of your research?

My favorite aspect of my research is that it connects advanced computational methods with real community needs. I am especially interested in how natural hazards affect transportation networks, because road closures can isolate communities, delay emergency response, and disproportionately impact underserved populations. Developing solutions to these challenges has the potential to directly benefit communities, and contributing to that process is one of the most meaningful parts of my work. I also value the collaborative nature of my research, where civil engineers, AI researchers, software developers, and transportation experts can exchange knowledge and build more practical, equitable solutions together.

What excites you the most about the opportunity to work on AI and Civil Engineering?

What excites me most is the opportunity to use AI not only as a modeling tool, but as a way to develop more robust, scalable, and actionable solutions for civil infrastructure problems. Natural hazards are complex, uncertain, and data-intensive, and AI can help us analyze large datasets, generate realistic hazard scenarios, quantify uncertainty, and support faster decision-making. For me, the intersection of AI and civil engineering is powerful because it can transform traditional infrastructure assessment into proactive, data-driven resilience planning.

What do you hope to accomplish during your time as a Chishiki AI Fellow?

During my time as a Chishiki AI Fellow, I hope to strengthen my ability to develop AI-based decision-support tools that are both technically rigorous and useful for real infrastructure planning. I am especially excited to collaborate with researchers in high-performance computing, AI, and civil engineering to improve the scalability of my framework. My goal is to use this fellowship to advance my research into a more robust platform for transportation risk assessment, and equitable infrastructure decision-making.

What do you like to do when you aren’t working on research?

When I am not working on research, I enjoy swimming and spending time with my family.

Auditable Engagement Design for Participation Bias in Transportation Planning

My research proposes an auditable AI framework for transportation planning that examines how public participation shapes infrastructure decisions. Transportation planning is not only a technical optimization problem; it is also shaped by who participates, how feedback is collected, and how trade-offs are communicated. My project uses GenAI-based stakeholder modeling, transportation performance evaluation, and large-scale simulation to compare different engagement mechanisms, including one-directional public comment collection, preference-weighted plan selection, and structured propose-critique-vote-revise deliberation. The goal is to make engagement rules themselves transparent and auditable, helping agencies understand when planning outcomes are robust to participation bias, when minority concerns may be overlooked, and how technical performance can be better aligned with collectively endorsed decisions.

Why did you become a scientist?

I became a scientist because I wanted to build rigorous computational tools for real-world transportation systems. My background in CEE showed me that transportation systems are shaped not only by roads, vehicles, and networks, but also by human behavior, incentives, policy, and uncertainty. This complexity drew me toward machine learning and decision-making methods that can help infrastructure systems become more reliable, adaptive, and equitable. I am especially motivated by research problems where technical advances can have direct societal impact, such as improving mobility, reducing emissions, and supporting more transparent public decision-making.

What is your favorite aspect of your research?

My favorite aspect of my research is working at the intersection of machine learning and transportation. I enjoy studying systems where algorithms must interact with people who have diverse preferences, incomplete information, and different levels of trust in technology. This makes the research technically challenging, but also impactful. In my current project, I am especially excited by the possibility of turning public engagement design into something that can be evaluated systematically with AI. By making participation bias, trade-offs, and decision rules more visible, I hope my research can contribute to transportation planning processes that are not only more efficient but also more accountable and inclusive.

What excites you the most about the opportunity to work on AI and Civil Engineering?

What excites me most is the opportunity to rethink civil engineering as a field where AI can support both technical optimization and better governance. Transportation systems affect everyone, but the processes used to design and operate them are often complex, uncertain, and difficult for the public to scrutinize. AI offers powerful tools for modeling, simulation, prediction, and decision support, but these tools must be designed carefully so they do not amplify existing biases or make decisions more opaque. I am excited by the chance to develop AI methods that are trustworthy, auditable, and grounded in practice. For transit planning in particular, AI can help evaluate multiple scenarios, reveal hidden trade-offs, and support more transparent conversations among agencies, stakeholders, and communities.

What do you hope to accomplish during your time as a Chishiki AI Fellow?

During my time as a Chishiki SCIPE AI Fellow, I hope to advance my auditable engagement framework for transportation planning and use advanced computing resources to test it at a much larger scale. The project requires many simulation runs across different transit planning scenarios, so access to strong computing resources will be essential to expand the scope and rigor of the research. I also hope to collaborate with researchers across AI, civil engineering, transportation, game theory, and decision systems to refine the framework and make it useful beyond a single case study.

What do you like to do when you aren’t working on research?

When I am not focused on research, I enjoy traveling, cooking, and spending time with friends and colleagues. Traveling gives me the chance to experience new places, cultures, and perspectives. Cooking is a creative and relaxing way for me to experiment, share food with others, and take a break from work. I also like learning from people across different fields and exploring new technical ideas outside of my immediate research. Recently, I have been especially excited about using GenAI to automate repetitive tasks. Whenever I find myself doing something more than once, I start thinking about how to turn it into a smarter and more efficient workflow.

Collaborative Spatial Intelligence in Multi-Robot Systems for Civil Infrastructure

My research advances collaborative multi-robot systems that achieve resilient autonomy in complex, dynamic environments. I develop algorithms that enable robot teams to coordinate navigation, manipulation, perception, and decision making in the real world. My goal is to build trustworthy and scalable multi-robot teams for civil, domestic, and safety-critical applications, including search and rescue, infrastructure inspection, and disaster response.

Why did you become a scientist?

I do not think I am fully a scientist yet, but one day I want to feel like one. For me, becoming a scientist means earning the ability to ask difficult questions and build something that can genuinely help society. I have always wanted my work to contribute something meaningful beyond myself, and robotics felt like the right field because it is both challenging and full of possibility. People are often skeptical about whether robots can work reliably in the real world, and that skepticism motivates me. I believe I can contribute to this field, and my goal is to help make robotics more useful, trustworthy, and impactful.

What is your favorite aspect of your research?

My favorite aspect of research is designing a system that can solve the problem. I enjoy this process because it depends on understanding the fundamentals, implementing the idea, analyzing the results, and then going back to improve the design. Research is never a straight path; it requires moving back and forth between theory, implementation, and evaluation. Among all of these steps, I find solution design the most exciting and critical, because it is where the main idea takes shape and becomes something that can eventually work in the real world.

What excites you the most about the opportunity to work on AI and Civil Engineering?

What excites me most about AI and Civil Engineering is the chance to work on problems that directly affect people and communities. AI and robotics can help with infrastructure inspection, disaster response, and search and rescue, especially in places that are risky or difficult for humans. I am excited by the possibility of building systems that are not only intelligent, but also useful in real world civil applications.

What do you hope to accomplish during your time as a Chishiki AI Fellow?

As a Chishiki AI Fellow, I hope to grow as a researcher and strengthen my work in AI and multi-robot systems. I want to develop algorithms that make robot teams more reliable, scalable, and useful in complex environments. I also hope to learn from others, collaborate across disciplines, and contribute to research that can have real societal impact.

What do you like to do when you aren’t working on research?

When I am not doing research, I usually like spending time by myself and staying at home. I enjoy relaxing, watching Netflix, playing games, going outdoors sometimes, and just thinking. Even when I am taking a break, my thoughts often somehow connect back to research, ideas, or problems I want to solve. For me, research naturally becomes part of how I see things.