Congratulations, Dean's Award Recipients!

I currently teach ECE 2214 (Physical Electronics) to sophomores and ECE 3214 (Semiconductor Device Fundamentals) to juniors and seniors, along with several graduate‑level semiconductor courses (ECE 6214, ECE 6204, and ECE 5205/5206).

I especially enjoy teaching device physics because it connects fundamental principles to real research challenges and industry needs.

I love helping students build confidence with challenging concepts. Some of my most meaningful moments come from students who later share how the course or my guidance made a difference in their academic path and, in some cases, even their lives.

Receiving the Dean’s Excellence in Teaching Award is an honor. It reinforces my commitment to high‑quality teaching and motivates me to continue supporting students with clarity, rigor, and care.

My research tackles a problem at the heart of modern software security: nearly every program we run — on our laptops, our cars, our medical devices, and beyond — ultimately executes as machine code, the raw instructions a processor understands. When a vulnerability is suspected in that machine code, we often don't have access to the original source code. In some sense, that is like trying to inspect a finished bridge for cracks without ever having seen the blueprints. Through a project I started at Virginia Tech called Low-Level Reasoning Machine (or LLRM), my group has been developing the foundations and tools to make that inspection rigorous, scalable, and trustworthy.

Several recent results illustrate the progress we have made. We have shown that a fundamental obstacle in the field — the impossibility of perfectly recovering instructions from a binary — can be tamed if compilers (the programs that convert source code to machine code) attach a small package of structured information alongside the binaries they produce: rich enough to enable rigorous analysis, yet compact enough to protect the original source (which is non-negotiable for closed-source software).

We have developed techniques that scale this kind of reasoning to programs of 400,000 to 500,000 instructions, the largest demonstrated to date, and that target memory safety, the class of flaws underlying many of the most damaging vulnerabilities in the wild. We have shown how to take an existing binary, repair or harden it, and redeploy it with mathematical, machine-checked guarantees that the modified version behaves exactly like the original everywhere except where we intended to change it. And we have built tools that treat the processor itself as the ultimate authority on what each instruction does, using systematic probing to recover instruction behavior more reliably than from manuals, which we have shown to contain errors and omissions.

These results — embodied in tools we have developed called ELLF, FoxDec, and libLISA — bring us closer to a future in which the binaries running on critical systems can be inspected, reasoned about, and repaired without access to source code and with the same rigor we expect in other engineering disciplines.

Even after some 350-plus papers, I still get a kick out of every paper acceptance — that part never gets old. But what I love most about research, and what gives it the deepest meaning for me, is watching my students grow into independent thinkers and problem solvers. The arc from a first-year PhD student to a mature scholar — navigating dead ends, persevering through paper rejections, reframing problems, finding their own voice — is the most rewarding thing I get to be part of. Having a front-row seat to those journeys is, honestly, an unbelievable privilege.

Although my name is on the award, this recognition really belongs to my PhD students, postdocs, and research faculty. They are the ones who do the hard, careful work, day in and day out. It is deeply gratifying to see that work get a pat on the back.

My recent research focuses on secure network infrastructures, such as DNS and 5G, and secure IoT (Internet of Things) systems. Most recent papers have been published in security and networking venues, such as USENIX Security'25, ACM CCS'25, and ACM MOBICOM'25, and NDSS'26. 

It is always fun to learn new things and gain new insights into a problem studied for a long time.  I enjoying seeing that the research results are helpful for other researchers and practitioners to resolve real issues in practice. 

This award is for my current and previous students and collaborators. Among them,  I am the least deserving of this award.

Recently, we have been working on reimagining how power electronics converters are designed, manufactured, and deployed within electrical systems. One approach we are developing is a modular concept that leverages the coaxial geometry and solid insulation of high-voltage cables. This approach enables power conversion anywhere there is room for a cable, and allows for flexible, scalable, and rapidly deployable electrical distribution for data center, grid, and transportation systems.

What I enjoy most about research is identifying new, unexplored problems and working at the intersection of real-world needs and deep technical challenges to develop meaningful solutions. I am especially drawn to the multidisciplinary nature of this work, where advances often emerge from integrating electrical, mechanical, and materials perspectives across scales—from materials and devices to circuits and systems. The most meaningful moments come from working with my students to turn ideas into working prototypes and watching them grow into confident researchers and problem solvers.

I am deeply honored to receive this recognition. It is especially meaningful to know that our work is valued and supported. This award reflects the dedication and hard work of my students, who are central to everything we accomplish. I plan to use the funds to support high school and undergraduate interns with hands-on research opportunities and to expand our lab’s capabilities with advanced equipment to further our mission in education, research, service, and outreach.

My research broadly focuses on trustworthy AI. I often say that trustworthy AI has to be built twice: once inside the model, and again in the system that puts the model to work. My lab studies both sides of that problem.

On the model side, we recently had a spotlight paper at ICML (accepted in the top 2.2% of submissions) where my student Bilgehan Sel, an Anthropic fellow, worked with Anthropic researchers to reveal a new vulnerability in Fine-Tuning as a Service models. We showed that frontier models can be fine-tuned to produce harmful outputs with very little “reasoning tax”, meaning the model retains its intelligence while becoming less safe.

On the systems side, I advocate for "System-First AI," where we evaluate AI based on the full deployed system: hardware, latency, cybersecurity, and human operators.

Recently, our collaborative team concluded a DOE project where we developed data-efficient AI methods to adapt to unseen cyberattacks, successfully deploying this technology with our utility partner, Southern Company, in a live power substation. Closing the loop from real system requirements to a working deployment in critical infrastructure was an incredible milestone.

What I love most is the freedom to explore important problems, guided by the goal of creating a real impact on society. We are in a period where AI is creating both remarkable opportunities and serious challenges across critical infrastructure and cybersecurity.

A particularly meaningful moment was seeing the AI technology our collaborative team developed actually deployed in a Southern Company power substation, actively monitoring for cybersecurity incidents. It was incredibly rewarding to see our utility partners go the extra mile to set that up with us, turning an industry-academic collaboration from research ideas into a real system that protects critical infrastructure.

Receiving this fellowship means a great deal to me because it shows that my work is supported at the college level and that I am part of a strong, collaborative ecosystem. I am immensely grateful for the mentors, ECE colleagues, students, and industry partners who make this work possible.

I plan to use the fellowship to directly support my students in their research and professional development. More broadly, it will help my group push the frontiers of trustworthy AI, building systems that are safer, more reliable, and aligned with the needs of society.

One of the things that I love the most about my job is meeting the students and building relationship. The thing that inspires me and motivates me every day is to see students be successful in reaching their goals, not just academic goals, but also their career goals, personal goals, and life goals. Whether it’s watching a student who struggled academically earn their Electrical or Computer Engineering degree, seeing an ROTC student commission to military service, or helping a student find their niche, at the end of the day, I am successful because my students are successful. And that's what motivates me every day - hoping that at some point during that day, I will have an impact, no matter how great or small, on a student and their success.

I am extremely humbled and honored to have received this award. A special thanks to Dan Stilwell, Virgilio Centeno and the Honorifics Committee for nominating me. I'm grateful to work in a department that recognizes the importance of our AP faculty and Advising Team - a team that offers support not only for our students, but for each other.