MIT hosts 11th Undergraduate Research Technology Conference
MIT students won 1st and 2nd place in URTC’s best paper awards
Elena Baskakova
From Oct. 10 to Oct. 12, the Stata Center was abuzz with bright minds and fresh faces as the Institute geared up for its 11th annual Undergraduate Research Technology Conference (URTC), where high school and undergraduate students from across the country came to present their latest research to experts and industry leaders.
Founded in 2015, the URTC is a collaboration between MIT and the Institute of Electrical and Electronic Engineers (IEEE). MIT has long been a member of the IEEE’s Boston chapter, which supports a variety of local universities in promoting engineering and research. However, for several years after its establishment, student engagement with MIT’s IEEE branch was concerningly low, especially among undergraduates.
After weeks of brainstorming possible ways to get undergraduates more excited about the club, MIT and the IEEE found the root of the problem. “The IEEE is [a] society that organizes a lot of technical conferences around the world,” explained Gim Soon Wan, Technical Program Chair of the IEEE Boston chapter. “Basically, they’re open to professionals, industrial professionals, graduate students, [and] PhD students.” However, as Wan and his team realized, it was nearly impossible for relatively inexperienced undergraduates to compete against such seasoned researchers.
Thus, the URTC was formed as a uniquely undergraduate-friendly conference, providing younger students with a pivotal opportunity to kickstart their careers in scientific research.
The conference has grown considerably since its conception, especially after allowing high schoolers to submit their own research papers. The URTC’s most successful year was 2025, where they received 319 submissions and accepted 189. This was a 40% increase in submissions from 2024, which held the previous record of 228 submissions and 154 acceptances.
From smart face masks and non-Newtonian fluid models to a robotic ASL instructor and an analysis of Spanish in Large Language Models, there was no shortage of innovations to witness.
An opportunity for high school seniors
According to high school senior Raunak Verma, the biggest challenge was the difficulty of adjusting to higher-level college research; this meant managing timelines, learning how to use professional language, and formatting in LaTeX. “There [was] a lot of all-nighters being pulled, tons of work, but it came together, and I'm really proud of our team,” Verma said.
Other high school seniors told The Tech that one of their major challenges was networking, especially when working with virtual resources. The act of contacting and coordinating with professors from different timezones was difficult, especially with many people working under them.
For these high school seniors, the most exciting part of the conference was the opportunity to engage with like-minded people — those oriented towards STEM — who might have similar interests. “Every other person you meet has been working on a really cool science project,” Verma explained; according to him, this made it easier to start conversations.
The opportunity to explore MIT was another highlight of their URTC experience, mainly for one reason: MIT is their dream school. As someone interested in robotics, Verma has always found MIT’s engineering culture fascinating, from interactive exhibits in the Media Lab to the rollercoaster built by East Campus during REX. During campus tours, Verma saw people in full hazmat suits working on nanotechnologies. “And I saw they made a chip smaller than rice, right? That's awesome,” he said.
Award-winning innovations
At the URTC awards ceremony on Oct. 11, five research papers were selected for a Best Paper Award.
First place: Turning up the heat
Jeewoo Kang ’27, Patrick Darmawi-Iskandar G, John Niroula ’18, SM ’24, PhD ’25, and Prof. Tomás Palacios from MIT were presented the Best Paper Award for their research paper titled “Current Collapse Measurements of SiN Passivated AlGaN/GaN HEMTs at Elevated Temperatures up to 500°C.”
The Tech interviewed Kang about her experience performing the research and presenting it at URTC.
Kang’s involvement in the project started with her inherent curiosity as a first-year at MIT. “I kind of came in knowing I wanted to do electrical engineering, but I didn’t really know what field I wanted to go [into],” Kang said.
She reasoned that gaining some experience through a UROP would help her decision. One day, she found a program that caught her attention — the Palacios group, which studies electronic and material innovations and works with transistors and metals. “This is stuff that a freshman wouldn’t really be able to experience [otherwise],” Kang said.
After completing training as a first-year, she officially joined the Palacios group as a sophomore and began her research into transistors.
“[A transistor is] just an electrical component,” Kang explained. “In a very, very simplified way, they act as a switch. So what we like to do is measure some current/voltage kind of characteristics of these transistors to see how well they fare.”
However, as Kang’s team discovered, most modern transistors have a critical limitation. “They’re typically made of silicon,” she said. “But silicon kind of deteriorates at certain temperatures. Anything above 200°C, it just doesn’t work as well.”
Kang’s task, therefore, was to research gallium nitride (GaN) as a more heat-resistant alternative to silicon in transistors. Her team investigated how well GaN-based transistors worked under temperatures up to 500°C.
Kang expects her team’s research to have various real-world applications in the near future, including in space exploration, the automotive industry, and other high-temperature applications. “I was hoping that we could see more gallium nitride or ultra-wide band gap semiconductors in these industries replacing the silicon and seeing how well they fare in actual application-based technology,” Kang said.
Second place: Simplifying a supercomputer
Elena Baskakova ’28, William Bergeron, Matthew Hubbell, Hayden Jananthan ’16, and Jeremy Kepner from MIT were presented the 2nd Place Best Paper Award for their research paper titled “TX-Digital Twin: Visualizing Supercomputer GPU Performance Data Stream.”
The Tech spoke with Baskakova about her research and experience at URTC. Baskakova also provided a visual presentation of her team’s work.
Baskakova’s involvement with this research project stemmed from her UROP with MIT SuperCloud and MIT Lincoln Laboratory Supercomputing Center, which form MIT’s supercomputer network. Baskakova helped monitor this supercomputer as part of her UROP.
“It turns out that this is actually a very non-trivial task,” she recounted. “A [super]computer is thousands of nodes, right? So it’s thousands of little computers, and each one has its own hardware, and it’s being shared amongst users, and they’re running jobs on the computers and so on.”
As Baskakova explained, keeping track of everything in this complex network of computers was crucial; a malfunction in an integrated system like that would be catastrophic.
Her team’s solution to this daunting task was to make it easier to visualize all the data they needed to monitor. The simulation they developed is a “digital twin” of the real supercomputer, taking in live data and using Unity to create a three-dimensional representation of it. Baskakova specifically handled designing the 3D graphics of the simulation and optimizing the simulation to make it run faster.
Baskakova later reflected on the joy this project brought her over the past several months. “As someone who is a designer at heart, I just really love making things for people,” she said. “So the research for me was more of, like, documentation, essentially, for what I did.”
Baskakova hopes that her team’s work helps inspire others in high-performance computing, from digital tone simulators to video game developers, to implement similar 3D visualization strategies.
Third place: How GenAI impacts misinformation
Saumya Chauhan, Mila Hong, and Maria Vazhaeparambil from Caltech received the 3rd Place Best Paper Award for their research paper titled “When GenAI Meets Fake News: Understanding Image Cascade Dynamics on Reddit.”
Their research focused on how generative AI and text leads to misinformation and how they both “contribute to virality of those posts within different communities, both at a post-level and then also cascade-level.”
According to Hong, what inspired the three to pursue this research project was the November 2024 presidential election. Seeing AI generated posts about then-presidential candidate Donald Trump on Reddit and X (formerly Twitter) made them concerned about how the Internet was prone to misinformation, especially via generative AI.
The trio said that after having completed their project, they would like to see regulations enacted, especially on Reddit, which is anonymous and lacks moderation compared to other social media platforms. However, according to them, it is difficult to enact policies at a federal level since that would “take away freedom of speech.” Instead, moderation can be done locally in online communities.
Hong found that the hardest part in executing their research was “getting more data.” Looking into Reddit posts, finding databases that “have image information itself” and creating “pseudo-labels for generative AI images” that may or may not be accurate were crucial, but hard to come by. Though there are annotated databases with AI images, they often lack the randomness of real AI-generated Reddit posts.
Nevertheless, Vazhaeparambil shared that “hearing very different perspectives” was one of the most exciting aspects of the conference and what made coming across the country from Pasadena all the way to Cambridge all the more worth it.
Fourth place: Giving a voice to those in need
Yujun Ge from Cerritos High School in California was presented the 4th Place Best Paper Award for her research paper titled “Transforming Assistive Communication: An AI and NLP-Powered Machine Learning Framework for Contextual AAC.”
The Tech interviewed Ge after the awards ceremony.
When asked how it felt to be at URTC that day, Ge spoke about how amazed she was by the passion and talent around her. “I feel really honored to be able to be surrounded by so many people that are interested in the things that they’re discovering,” she said.
Ge’s research paper topic was rooted in her experience as a tennis coach for players with autism. According to Ge, non-verbal people with autism have difficulty communicating, even through the current Augmentative and Alternative Communication (AAC) system most use. Many find it burdensome, as the system is too robotic and lacks customization.
“My research [creates] an adaptive and really personalized system to help these people speak honestly, giving a voice to them again by [providing] them [with] adaptive phrase suggestions in real time,” Ge said.
Ge’s communication system analyzes environmental cues, such as the time of day, day of the week, location, and what is being said nearby, to generate phrases for the user to respond instantly with during conversations. This process cuts down the time it takes to form a sentence from minutes to a couple seconds.
Ge is adamant about making her innovation a real product. One of her core values is not just doing research on paper, but “bringing it out to the market, to the people who can actually benefit [from] it.”
This principle is part of the reason Ge pursues research in general. She emphasized that she never does research for personal gain or a boost on her college applications; rather, she has always been focused on helping people that are often overlooked.
“I want to choose technology for humanity,” Ge said.
Fifth Place: Automating Scientific Breakthroughs with AI
The 5th Place Best Paper Award went to high school senior and Research Science Institute alumnus Aditya Sengupta from the Kellis Lab at MIT for his research paper titled “SPHINX: An Agentic AI System for Automated Scientific Discovery in Computational Biology.”
Research Science Institute (RSI) is a six-week long summer program held at MIT where 100 students from different states and countries are paired with Institute researchers to work on a STEM project. During RSI, Sengupta collaborated with Professor Manolis Kellis of the Computational Biology Group at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL).
For Sengupta, what struck him the most about MIT URTC was “the energy of seeing peers tackle challenges” in various fields, from robotics to biomedical engineering. Being around hundreds of people also passionate about science – “a community using science to solve some of the world's most pressing problems,” in Sengupta’s words – reminded him about why he entered research in the first place.
SPHINX is an agentic AI system designed to automate scientific discovery in computational biology. Often, in computational biology studies, large amounts of data are collected and published, from RNA sequencing to protein structures. However, according to Sengupta,“Extracting actionable insights remains bottlenecked by manual workflows.”
Sengupta’s solution to this pressing problem is SPHINX. Given a high-level scientific question, SPHINX delegates tasks to six sub-agents, which, according to Sengupta, specialize in tasks such as “literature synthesis and hypothesis generation, medical knowledge validation, intelligent dataset curation, data analysis and visualization, in-silico experimentation using ML models, and synthesizing conclusions and wet-lab experiments.”
With access to approximately 200 tools, including databases and virtual cell models, SPHINX was able to generate, as Sengupta demonstrated, “Nature journal-quality published findings from Alzheimer’s disease research.” Moreover, using computational models, SPHINX was also able to predict how immune cells would react with drugs. Essentially, as Sengupta explained, the system does the repetitive tasks, “freeing human scientists to focus on high-level design and interpretation.”
The project was inspired by the 2024 Nobel Prize in Chemistry, which was given to David Baker, Demis Hassabis, and John Jumper for “computational protein design” and “protein structure prediction,” according to a press release from The Royal Swedish Academy of Sciences. This award showed Sengupta that AI was “a technology that has the power to transform how we do science itself.”
When asked about his project’s applications to the real world, Sengupta outlined that, first, SPHINX can be used to accelerate discovery timelines – such as those for drug discovery and disease modeling – and reduce costs. “SPHINX can reduce the time from data collection to validated hypotheses and results by months or years and allow faster treatments and cures for patients,” he said.
Seungupta also talked about how SPHINX can be leveraged by “smaller labs, institutions in underserved regions, and domain researchers” to engage in research without having to go through extensive training in AI workflows.
Besides increasing accessibility in research, SPHINX may be able to open new scientific frontiers. Sengupta explained that not only would “researchers gain bandwidth to tackle deeper questions at a higher level of creativity,” but they could also “uncover new patterns” in biological data that human intuition might miss, leading to new breakthroughs.
The future of URTC
Conference advisor Jie Wang, a postdoctoral research associate at MIT, said that meeting high school seniors and undergraduate students from different corners of the world and learning about their projects is what makes URTC exciting. According to him, the younger generation has a lot of potential to positively impact the world by learning and applying “advanced technologies and sciences to different problems.”
The biggest challenge of organizing URTC this year was the number of people attending the event; on top of that, the conference had people from different states and countries from New Zealand to South Korea. URTC has been running for more than a decade, but it has never seen turnout like this. For Wang, when it came to dealing with such a large and diverse amount of people, the main questions were “How to [help attendees] learn in the conference?” and “How [to make attendees] enjoy the conference?” Wang hopes that URTC has contributed to increasing the exchange of ideas and cultures across national boundaries.
Though the increased attendance was perhaps the biggest challenge that organizers faced this year, Wang expressed a desire to expand URTC nationwide. His vision for the conference is one in which there is a flagship event to provide high school students and undergraduates with “more access and learning opportunities.”
Wan said that word of mouth triumphs over every other form of marketing out there. “We did not spend any single cent on [advertising] or publishing,” he said. Instead, he distributes merchandise, allowing attendees to talk about the event with friends and family. By doing so, people would become intrigued by the conference, leading them to begin “thinking about doing research [and] writing [a] paper to submit to this conference.”
Wan said that the satisfaction from helping run URTC not only comes from giving high schoolers the opportunity to present their work and engage in STEM activities to “nurture future researchers,” but also allowing them to explore MIT and “see what it’s all about.”
With the number of attendees increasing and plans to expand the conference further, URTC organizers expect that it will become a leading STEM conference for high school students and undergraduates, allowing more student researchers to celebrate their efforts and engage in a community of like-minded peers.