Duran spent nearly four months interning at Apple’s headquarters in Cupertino, California. His work during the internship focused on Apple’s Foundation Models, a series of generative AI models that power Apple Intelligence, and on the large-scale infrastructure and tooling around them.
Duran’s research interests lie at the intersection of computer architecture, systems, and machine learning. More recently, he has focused on making artificial intelligence (AI) workloads more accessible, sustainable, and cost-effective. Duran is co-advised by Associate Professor Yingyan (Celine) Lin and Assistant Professor Anand Iyer in the School of Computer Science.
Duran said he saw the internship as an opportunity to gain industry experience and apply his knowledge to real-world problems.
“Industry research is particularly exciting to me because you can work on challenging problems while also seeing how these ideas can translate into systems and products at a very large scale and impact nearly a billion lives,” he said.
As a research intern, Duran especially appreciated the level of independence he was given. He was expected to take initiative to find problems on his own and then develop solutions with his mentors. Although Duran liked taking ownership of his work, he also came to better understand the value of outside input.
“One thing I’m continuing to improve is recognizing when it’s more efficient to ask for input earlier rather than spending too much time trying to solve everything independently,” he said.
Duran said he enjoyed the internship, especially the people he met. However, the biggest highlight for him was presenting his work.
“I had the opportunity to present several times, including to larger audiences across different organizations, and I really enjoyed discussing the findings and getting feedback from people with different perspectives,” he said.
In addition to the technical knowledge he gained, Duran said the experience gave him a new perspective on how research and engineering work in industry.
“I realized that combining strong research with real business needs can often create much greater impact than research alone,” he said.
“I really enjoyed seeing my work translate into something with broader, tangible impact, and that has influenced the kind of work I want to pursue in the future.”
]]>The kitchen serves students in need with groceries or meals, works to reduce food waste on campus, and is part of Students’ Temporary Assistance and Resources (STAR) Services.
“The STAR Office is so grateful to the School of Cybersecurity and Privacy for sponsoring a food drive this semester,” said STAR Program Director Steven Fazenbaker. “Your successful campaign has put us in a position to serve over 300 students by keeping Klemis Kitchen fully stocked with our most popular shelf-stable food items for the rest of the fall semester.”
“It also allows us to get a head start restocking Klemis Kitchen during the winter break, so we will be ready when the spring semester begins. Thank you so much for your over-the-top generosity.”
The school exceeded its goal by $1,600 and surpassed last year’s donation of $2,110.
Director of Research Operations Mary Helen Hayes organized the food drive and spent September fundraising.
“While I organized and led the drive, its success came from the generosity and community spirit of our faculty and staff,” Hayes said.
“Thank you to everyone who contributed and helped make it possible. SCP is proud to support such a meaningful cause, and we truly hope our contribution makes a difference for Klemis Kitchen and the students it serves.”
While SCP hosts this food drive once a year, Klemis Kitchen accepts donations year-round.
Visit Klemis Kitchen to learn more about how you can help. The Kitchen relies 100% on donations, including leftovers from the dining halls, donations from community partners like grocery stores and churches, and food drives sponsored by departments across campus.
Photos by John Popham/College of Computing
]]>In a study accepted for the 2026 ACM Conference on Computer and Communications Security, Georgia Tech researchers interviewed 20 military sailors about their cybersecurity experiences, training, and responses to hypothetical scenarios. Their findings point to a gap between how mariners are trained to handle emergencies and how cyberattacks can compromise the systems they depend on.
The study found several gaps in how crews prepare for cyber threats:
These gaps matter because cyberattacks can compromise a system without causing it to stop working. For example, an attacker could manipulate navigation data, leaving a crew to make decisions based on false information.
“I’d assume it was anything but cyber at first. It would probably take me quite some time to get to that conclusion,” one participant told the researchers.
The team examined three hypothetical scenarios: GPS spoofing, loss of propulsion control in a high-traffic area, and failure of electronic navigation systems in a congested channel at night.
In each scenario, mariners would first focus on keeping the vessel safe, using established procedures to stabilize the ship and troubleshoot its systems. While this response is essential during an emergency, it could delay recognition of an attack if crews do not consider that the system or its data may have been compromised.
For example, a ship navigating a narrow channel depends on accurate location information, but what happens when its navigation systems appear to work normally while a hacker manipulates that data?
An attacker could use GPS spoofing to falsify a vessel’s reported position. If the false position is close enough to the ship’s actual location, a mariner might not notice anything wrong. The vessel could gradually head toward shallow water or another obstacle without triggering an obvious error or alarm.
“By the time the mariner realizes something is wrong, the ship’s size and limited maneuverability could make it too late to avoid a grounding or collision,” said Anna Raymaker, a Georgia Tech Ph.D. student who studies maritime security.
The scenario illustrates a central challenge in maritime cybersecurity: An equipment failure may be easy to recognize, but a cyberattack that provides false information can be much harder to detect.
Batten the Hatches: Cybersecurity with Military Mariners was written by master’s student Ryan Von Brock, Raymaker, Research Scientist Animesh Chhotaray, Associate Professors Frank Li and Saman Zonouz, and Provost & Executive Vice President for Academic Affairs Raheem Beyah.
The research was conducted in Georgia Tech’s Cyber-Physical Security Lab, a collaboration between the School of Cybersecurity and Privacy and the School of Electrical and Computer Engineering.
Photos by Terence Rushin/College of Computing
]]>Cognitive behavioral therapies require patients to practice skills between sessions. Clinicians rely on retrospective patient reports to understand what happens during that time.
However, recall bias, selective disclosure, and difficulty remembering can render those reports subjective, incomplete, or inaccurate.
Clinicians can now see sensor data on how patients’ bodies react in these situations thanks to a new technology toolkit from Georgia Tech and Emory University.
A new app that syncs with a FitBit wearable device can track and record heart rate, step count, noise level, and other sensor data when patients practice prolonged exposure homework assigned by their therapists.
A web dashboard lets clinicians review that data, add notes, and track patient progress. This information helps them structure therapy sessions more effectively and accelerate patient progress.
The toolkit is the culmination of a $1.2 million National Science Foundation grant awarded to Georgia Tech’s Rosa Arriaga and Emory’s Andrew Sherrill in 2019.
Arriaga is an associate professor at Tech’s School of Interactive Computing. Sherrill is an assistant professor in the Department of Psychiatry and Behavioral Sciences at the Emory School of Medicine.
Sherrill is also a practicing clinical psychologist at the Emory Healthcare Veterans Program and has worked with veterans with PTSD for more than a decade.
Arriaga and the students in her Ubicomp Health and Wellness Lab spent five years developing the toolkit, which they call the Patient Engagement Computational Sensing System (PECSS).
“It was a lot of master’s students and undergrads who built the back end to the app and the dashboard,” Arriaga said. “It was a lot of work and effort by a lot of students to take an idea from conception to deployment.”
The researchers originally titled the toolkit the Prolonged Exposure Collecting Sensing System, but they changed the name after realizing it could be used for other types of therapy.
“The idea is that for any therapy that has a homework component, this can bridge the gap between patient self-reporting and clinician intuition,” Arriaga said.
Nathaniel Swinger, a Ph.D. student in Arriaga’s lab, co-led a paper on the study’s results along with Ph.D. student Myeonghan Ryu. Their paper received a best paper award at the 2026 ACM Conference on Computer-Supported Cooperative Work and Social Computing (CSCW). Best papers are given to less than 1% of the papers published at the conference.
“Our findings showed that the deployment was quite successful,” Swinger said. “Over 75% of the patients who incorporated PECSS into their treatment experienced a clinically significant change in symptoms. We don’t know if that’s causal, but there is some evidence the system worked to facilitate the process.”
Prolonged exposure is a method for systematically approaching trauma memories and reminders to make them safe and tolerable for patients. It is typically delivered in one-on-one therapy sessions.
Sherrill said the toolkit shows promise toward enhancing prolonged exposure therapy outside of clinical settings.
“Prolonged exposure has been researched more than any other therapy that exists,” Sherrill said. “We know it’s effective, but we’ve had problems for decades implementing it in the real world.
“There’s a need for technology in taking what patients learn in their sessions and having them replicate it outside of the clinic.”
Until now, the impact prolonged exposure on a patient with PTSD depends on the patient’s willingness to self-report the results of their homework.
“The number of potential triggers is an endless list,” Sherrill said. “We want people to approach them with their guard down and not to engage in avoidance or soothing behaviors.”
Sherrill said most of his patients have shown interest in continuing to use the toolkit — even after therapy is no longer needed.
“A system like this could reduce the frequency of sessions with a therapist,” Sherrill said. “We’re teaching patients to be their own therapists. This toolkit gives them a greater sense of independence.”
Given the promising results, Swinger said that ideally the next step would be a large-scale clinical trial.
“We’re looking to better test the causal clinical impacts of the system and a more generalizable notion of efficacy,” Swinger said. “We want to rigorously test it in a clinical setting for people beyond veterans with PTSD.”
]]>New research from Georgia Tech suggests that a simple change to a chatbot’s instructions could reduce that risk. The work was accepted to the ACM AI Letters publication.
Researchers Lorenzo de la Loza and Vijay Madisetti tested eight AI models in 1,280 experiments involving simulated conversations about delusional beliefs. They found that instructions designed to discourage excessive agreement reduced the models’ confirmation of those beliefs by 73.8%. Adding a self-reflection step increased the reduction to 77%.
“We think that this paper will help AI providers and solution creators better understand how best to protect and safeguard their users from harming themselves,” Madisetti said. “This may make AI safer for all of us.”
AI chatbots are designed to be helpful and supportive. But that behavior can become risky when a chatbot agrees with a user’s beliefs without questioning whether they reflect reality. Researchers call this behavior sycophancy.
Courtney Crooks, a principal research scientist at the Georgia Tech Research Institute and adjunct faculty member in Georgia Tech’s School of Cybersecurity and Privacy, said chatbot conversations can have both positive and negative effects.
People may feel more comfortable discussing difficult thoughts with a chatbot than with another person. But if the chatbot always agrees, users may miss opportunities to question their beliefs or consider other points of view.
The same behavior that makes a chatbot feel supportive, Crooks explained, can also encourage harmful patterns of thinking.
The researchers tested instructions that told AI models not to reinforce users’ beliefs simply to be agreeable. The team evaluated the models using 16 clinically derived scenarios involving delusional beliefs.
The team also tested other safety tools, including systems designed to flag potentially harmful content. The results suggest that changing a chatbot’s instructions can be an effective way to reduce excessive agreement in the scenarios tested.
Crooks called the findings a promising direction for future research combining psychology and AI safety.
The study did not test people in clinical settings or prove that the approach prevents real-world harm. However, it offers developers a potential way to make chatbots more careful when responding to vulnerable users without building an entirely new AI model.
]]>Communications Officer for the School of Cybersecurity and Privacy
]]>A team of computer scientists and doctors led by Georgia Tech aims to change that. Using a system of artificial intelligence (AI) models, the team can help clinicians deliver the right intervention to the right patient at the right time.
Their system uses Restless Multi-Armed Bandit (RMAB) models. While the name is technical, the idea is simple. The RMAB framework helps decision-makers choose the best actions over time when conditions are complex and constantly changing.
This project will tackle Type 2 Diabetes (T2D), building on previous experience expanding RMAB models to support maternal health care in India. Success on this front could lead to treating other chronic illnesses, like hypertension, heart disease, kidney disease, and mental health.
Recognizing the RMAB system’s potential, the National Institutes of Health (NIH) awarded the group an R01 grant. The award will fund the project with $1.2 million through 2029.
“Our goal is to address the immense challenge of allocating limited health intervention resources in optimal ways,” said Kai Wang, an assistant professor in the School of Computational Science and Engineering (CSE).
“By leveraging rich contextual data and accounting for social determinants, we envision RMAB models that empower health centers to customize interventions to individual patient profiles. Ultimately, this advances precision medicine and ensures equitable healthcare access for diverse populations.”
T2D is a diabetes condition where the body resists insulin or doesn’t produce enough to regulate blood sugar levels. Left unmanaged, high blood sugar leads to severe health problems like heart disease and stroke, damage to organs like eyes and kidneys, nerve damage, infections, and ulcers.
T2D is challenging to treat because patients have different ages, genetics, lifestyles, and body composition, all of which affect the progress of the disease. Social factors, like transportation and family support, matter just as much as biology to treat the disease.
For these reasons, patients who get the same medications may get very different results. While treatment plans now incorporate new GLP-1 drugs, they remain expensive and difficult to distribute equitably.
“Diabetes is a great example of a complicated, multi-faceted problem. It is a natural playground for the kind of high-dimensional machine learning algorithms we want to build and deploy,” said School of CSE Professor B. Aditya Prakash.
“As a chronic disease, diabetes also creates opportunities for sequential decision-making. Providers are not making just one decision to treat the condition, but rather a series of decisions over time.”
True to its multi-armed name, the team’s framework applies RMAB models across three different thrusts.
Contextual RMABs use patient histories, biomarkers, demographics, and other data to predict how each individual will respond to an intervention. These models directly provide personalized, data-driven treatments.
By incorporating patient-specific information, contextual RMAB models can help healthcare providers prioritize the most timely interventions. The goal is to improve patient outcomes while making more efficient use of limited healthcare resources.
Next, the group wants to understand social life and community influence in public health. Families and social networks shape how people manage chronic health conditions.
Network RMABs study and account for the role of social factors in health interventions. For example, the models can recognize when one patient responds better to treatment than others through benefits of their social circle.
More broadly, hospitals, clinics, government agencies, and nonprofit organizations often share responsibility for the same population in an area. Sometimes, they even compete for the same medications or funding.
Multi-agent RMABs are the third glass of models aimed at decentralized health systems to make fairer decisions. These tools examine how to coordinate resources and decisions between stakeholders. This in turn leads to policies that encourage cooperation and equitable resource distribution.
“I feel extremely confident in RMABs because we have already seen these frameworks deliver real-world impact,” Wang said.
“Prior success has proven the value of data-driven decision-making under resource constraints, giving us a strong blueprint to effectively address the nuances and complexities of diabetes.”
Wang previously built RMAB models to expand maternal healthcare for women in India. Working in conjunction with the non-profit organization ARMMAN, the models used healthcare information to identify and reach the most vulnerable populations.
By integrating RMAB models into ARMMAN mobile health technology, the partnership has reached over 350,000 mothers. This further led to a 31% improvement in engagement with health information.
Compared to maternal healthcare, diabetes is a more nuanced disease with differing causes, symptoms, and treatments. This requires a team effort spanning institutions and backgrounds.
Dr. Sara Cromer is an endocrinologist at Massachusetts General Hospital (MGH). Experienced diagnosing and treating diabetes, she has studied how social issues and equity influences the disease. MGH will provide much of the patient data to train the RMAB models.
Associate Professor Jian Li studies reinforcement learning at Stony Brook University. His expertise in applying reinforcement learning for resource allocation plays a key part tailoring models for the T2D problem. Wang and Li have previously worked together on RMAB research.
Prakash is an expert in computational epidemiology at Georgia Tech. He has led substantial contributions to the use of AI for networks to predict outbreaks and track disease spread. This experience has led to collaborations with NIH, the Centers for Disease Control and Prevention (CDC), National Science Foundation, and others.
“This project speaks to the interdisciplinary problem-solving we try to achieve in CSE. Despite having different backgrounds, our interest in applying our strengths to real-world healthcare problems unites us together,” Prakash said.
“These problems require experts from all sides. Without computational scientists, we can’t take advantage of available data. We lean on clinical experts to provide the data and expertise needed to guide our methods and make sense of results. Ultimately, these approaches need to be clinically validated if we want to make real-world progress.”
]]>A team of computer scientists and doctors led by Georgia Tech aims to change that. Using a system of artificial intelligence (AI) models, the team can help clinicians deliver the right intervention to the right patient at the right time.
Their system uses Restless Multi-Armed Bandit (RMAB) models. While the name is technical, the idea is simple. The RMAB framework helps decision-makers choose the best actions over time when conditions are complex and constantly changing.
This project will tackle Type 2 Diabetes (T2D), building on previous experience expanding RMAB models to support maternal health care in India. Success on this front could lead to treating other chronic illnesses, like hypertension, heart disease, kidney disease, and mental health.
]]>After a grueling, world-class competition, the team won the $4 million first-place prize in August 2025.
The team has since disbanded. Some members have moved into the corporate world, while others continue in academia. Despite these varied paths, Team Atlanta’s search for breakthrough AI cybersecurity solutions continues.
Team Atlanta lead Taesoo Kim, principal security researcher Dongkwan Kim, and three other former team members joined Microsoft Security after the competition, where they continue to work on AI and cybersecurity in the Frontier Offensive Research & Generative Exploitation (FORGE) Lab.
The move has given the researchers an opportunity to apply concepts from AIxCC to real-world software. It has also exposed them to challenges that are difficult to replicate in a competition or research lab.
Dongkwan Kim said that team members continue to share what they learned through research papers, reports, software, and benchmarks. At Microsoft, these lessons are tested against real customer code.
This work led to MDASH, Microsoft's multi-agent, multi-model system for vulnerability discovery and remediation. The system uses more than 100 AI agents to search for vulnerabilities, validate potential findings, and help develop fixes.
Rather than asking a single AI model if a piece of code contains a vulnerability, MDASH can examine a problem using multiple agents. The system can also try to create an exploit to determine whether a potential vulnerability is real and reachable.
For Dongkwan Kim and the other former Team Atlanta members at Microsoft, moving from AIxCC to MDASH showed how research can look different when applied to real software.
“AIxCC was not an endpoint for Team Atlanta,” he said. “We continue sharing what we learned through technical reports, papers, open artifacts, and benchmarks for vulnerability discovery and patching.”
Working with customer code introduced new challenges for the Microsoft team. The researchers had to consider security boundaries, existing development processes, how findings would be reviewed, and how vulnerabilities would be prioritized for action.
These challenges created a new feedback loop between research and product development. The team took ideas from AIxCC, test them on real software, learn where they worked and where they fell short, and use those lessons to improve the system.
The experience also showed that AI does not replace the need for cybersecurity experts. Human researchers are still needed to provide context, decide how to break down difficult problems, validate findings, understand attack paths, and determine the potential impact of a vulnerability.
Taesoo Kim, a professor in the School of Cybersecurity and Privacy and vice president of security research at Microsoft, said the competition continues to influence the Microsoft team's work.
“Every single activity that we’ve done at Microsoft was influenced by AIxCC,” he said in a 2026 interview with Microsoft.
The work has continued to grow. In 2026, Microsoft introduced MAI-Cyber-1-Flash, an AI model built specifically for cybersecurity and designed to work with MDASH.
The model is designed to handle many of MDASH's tasks while allowing larger AI models to focus on harder problems. Microsoft reported that replacing 80% of the models used by MDASH with MAI-Cyber-1-Flash increased its performance on the CyberGym benchmark from 88.4% to 95.95%.
The benchmark includes more than 1,500 examples from 188 software projects. Microsoft also reported that the new setup can achieve similar performance at half the cost of the previous MDASH configuration.
“AIxCC has sparked cybersecurity innovation across the globe, from frontier AI companies to small academic labs,” said Taesoo Kim. “It is an incredibly exciting time to be working in AI-driven security engineering and research, as we see AI fundamentally reshape how we find and fix vulnerabilities at scale.”
The AIxCC experience also continues to shape cybersecurity research at Georgia Tech.
Jiho Kim, a Georgia Tech Ph.D. student who was part of Team Atlanta's research efforts, said one of the biggest lessons he’s brought back to campus was learning how AI and traditional security tools can work together.
One of these tools is fuzzing. Fuzzing tests software with large numbers of inputs to detect unexpected behavior or security problems. Large language models (LLMs), meanwhile, can reason about code and help guide the search.
During the AIxCC competition, the Team Atlanta combined these ideas in CtxFuzz, a system that uses an LLM to help guide fuzzing toward parts of software that may contain vulnerabilities.
The research was submitted to the ACM/IEEE International Conference on Automated Software Engineering, where it received the ACM Distinguished Paper Award.
“Since AIxCC, we’ve continued building on some of the core findings from our two-year journey with Team Atlanta,” Jiho Kim said.
“One of the biggest lessons for us was understanding how LLMs and traditional fuzzing can work together in a mutually beneficial way, using each to overcome the limitations of the other and ultimately find more and deeper vulnerabilities.”
The researchers also wanted to make the lessons from AIxCC useful to other cybersecurity researchers.
They created an Open-Source Cyber Reasoning System Framework (OSS-CRS) and CRSBench, which provides researchers and security professionals with ways to use, test, and build on technology developed through AIxCC.
“We’ve also been focused on making the results of AIxCC useful to the broader security community,” Kim said. “We created OSS-CRS and CRSBench so that researchers and practitioners can more easily use, reproduce, and build upon the capabilities and lessons from our AIxCC work.”
The team's research has also moved beyond identifying vulnerabilities. Researchers are studying how well AI models can understand and reason about computer code.
Jiho Kim and his lab mates in the Systems Software & Security Lab developed COIN, a benchmark designed to test how well LLMs can reason about code. The work has been accepted to the 2026 Conference on Empirical Methods in Natural Language Processing.
The Georgia Tech researchers are also developing additional benchmarks to study the various types of reasoning required for cybersecurity. They are using what they learn to create new datasets and methods for training AI models focused on cybersecurity.
“We’re using what we learn from these benchmarks to train cybersecurity-focused models through new datasets and training methodologies,” Jiho Kim said.
Photos by John Popham/College of Computing
]]>Communications Officer at the School of Cybersecurity and Privacy
]]>