A team of Purdue University students, including Purdue Polytechnic senior Emma Mahan, earned third place out of 80 international competitors in the 2025 TE-AI Cup, an annual artificial intelligence competition hosted by TE Connectivity.
Alisha Willett reports that the four-member team, which also ranked second among all U.S. universities, developed an innovative AI model to detect and diagnose faults in centrifugal pumps. Their breakthrough approach proved effective even under conditions the system had not previously encountered.
Mahan, a student in the computer and information technology (CIT) program, was a key member of the interdisciplinary team. She explained how her hands-on, applied education from the Polytechnic was crucial to the project's success.
"I believe my IT background highly strengthened my abilities in data engineering and preprocessing, which I used to set up robust pipelines to handle complex sensor data efficiently," Mahan said. "I also utilized my experience in system integration to ensure that our model integrated smoothly with existing infrastructure and hardware systems."
The team also included students from Purdue’s Department of Computer Science and Department of Statistics. Mahan noted that this blend of different academic backgrounds was a major advantage, allowing each member to contribute a unique perspective.
"The most valuable part of working in a cross-disciplinary team is learning about different perspectives and approaches to the same task," Mahan said. "For example, working with computer science majors meant they were focused on code optimization and modularity. As a CIT student, I worked on looking at the overall picture from the business perspective, making sure we had the right deliverables for our stakeholders."
The project was organized through the Machine Learning @ Purdue student organization, and the team used Purdue’s high-performance computing cluster to train their model. In recognition of their achievement, the team received a $2,500 scholarship from TE Connectivity.
Additional information
Read the original reporting from Purdue’s Department of Statistics