Courtesy of Vihaan Bhaduri
Bhaduri’s test drones.
Whether in the air or in the operating room, juniors Vihaan Bhaduri and Ansh Chaudhary are using their love for technology to push boundaries. While Bhaduri has been working with German researchers to improve drone stability and autonomous flight, Chaudhary is busy developing AI tools to analyze CT scans and support surgical planning. Though their projects are in very different fields, both students are united by a shared mission: using AI to make complex systems safer, smarter, and more efficient.
Since middle school, Bhaduri has been fascinated by drones and flight technology. Now, he’s published a research paper analyzing drift in prototypical nano quadcopters, and has since opened the door to an internship with the German research agency RT Lola.
His interest in drones was sparked during a family trip to Peru in 2023, where he saw researchers use drones in the Amazon to monitor rainforest biodiversity. Since then, his curiosity has evolved into a technical and academic exploration of flight safety and drone autonomy.
“The entire world is shifting toward autonomous systems,” Bhaduri said. “Drones are a big part of that, from deliveries to defense. Making flight more reliable and cost-effective is going to matter more as drones become widely used.”
Bhaduri’s project originally began with him comparing manual drift measurements through drone flight tests with data from the onboard logging system in Crazyflie drones. His results revealed that the drones’ built-in tracking tools were often unreliable. Since then, Bhaduri has shifted to working with RT Lola to integrate real-time monitoring into miniature drones to help tackle this issue.
The collaboration has allowed him to explore auto-correcting flight paths for autonomous navigation. Bhaduri conducts most of his drone testing inside a netted flight cage in his garage, which he converted into a makeshift drone lab with the help of his dad. Testing in a controlled environment helps him avoid costly crashes; the drone parts, which are shipped from Europe, carry high price tags.
Though his current project is collaborative, his previous research paper was completed with guidance from a professor from Barnard College whom he worked with during the pandemic. While his research is demanding — taking as much as 30-40 hours each week — Bhaduri said it has helped shape his career ambitions. He’s still deciding what exact field to pursue, but he knows it will involve AI and flight in some capacity.
Chaudhary’s commitment to AI in medicine
Chaudhary’s research focused on adapting a cutting-edge transformer model — originally developed for natural language processing — to analyze CT scans. The goal of his research is to help surgeons identify organs more accurately by using AI to detect anatomical structures in the scans.
While working on his project, Chaudhary originally aimed to train his model using 2D medical scan data that clearly outlined different organs. However, he quickly discovered that there was very little publicly available 2D data. Instead, he found a 3D CT scan dataset.
To adapt it to his needs, Chaudhary manually converted the 3D scans into 2D slices by cutting them into individual cross-sections. He then used these slices to create his own custom dataset for training the model. Chaudhary said sourcing the right kind of CT scan data proved to be one of the most difficult parts of the process due to privacy restrictions and limited availability.
“I had to go through full 3D scans and manually slice them into hundreds of 2D images — it took hours of careful work just to build a dataset the model I could actually use,” Chaudhary said.
To evaluate the AI’s accuracy, he used common computer vision metrics like precision to compare how accurately the model could highlight organ structures within the images. The results showed promising potential for surgical use, where doctors could visualize organ locations more clearly before making incisions.
Throughout his research, Chaudhary was supported by a mentor from Cambridge University, who helped guide him through the paper-writing process. He eventually presented his findings in person at a conference hosted by the university, where professors in the research field around the world offered feedback on his work.
While the project confirmed Chaudhary’s interest in AI, it also helped him realize he may not want to pursue medicine professionally, since it did not interest him as much as the research did. Instead, he sees himself continuing in AI work, finding applications in various fields to make a meaningful impact.
Both students say that despite the steep learning curves and time commitment, their experiences have given them a clearer picture of the fields they want to pursue — and the confidence to keep innovating. They hope to continue building on their research in college, where they plan to dive even deeper into the intersection of technology, science and real-world problem solving. For now, both are focused on refining their skills — one flight and one scan at a time.
