Senior Design Spotlight: Using Machine Learning to Improve Pump Reliability

As an industry-partnered senior design project, the BD Infusion Pump team is working with medical device company Becton, Dickinson and Company (BD) to tackle a real-world engineering problem with the potential to impact millions of patients. Using an interdisciplinary approach, the team is developing a machine learning model to predict degradation in Alaris infusion pumps through inrush current analysis.

The cross-disciplinary team includes Jackson Downey '26 (IntE), Julia Marie Greely '26 (IntE), Sean O’Keefe '26 (IntE), Caley Tamondong '26 (IntE), Phillip Banky '26 (CS) and Jennaya Horne '26 (CS). The project brings together biomedical engineering, computer science and data analysis, allowing the team to connect hands-on hardware testing with data-driven modeling. 

The integrated engineering (IntE) biomedical engineering subteam — Downey, Greely, O’Keefe and Tamondong — has focused on experimental testing and dataset development, collecting inrush current data from both healthy and intentionally degraded infusion pump components at BD’s Pacific Mesa campus.

The computer science subteam — Banky and Horne — has worked closely to understand the device’s system architecture and develop a machine learning model best suited to the problem.
  Jennaya Horne (left), Caley Tamondong (center) and Julia Marie Greely (right) present the team’s semester progress, including project planning, workflow design and a custom data acquisition circuit

Throughout the project, the team has worked closely with Dr. Bradley Feiger, an engineer at BD, who serves as the team’s primary industry mentor. With expertise in medical device engineering and AI-related software development, Dr. Fieger provides technical guidance, feedback on design decisions, and insight into real world medical device research and development.

This fall semester has been shaped by close collaboration with industry stakeholders and consistent iteration. The team has spent time digging into medical device documentation, interviewing engineers across disciplines and building on prior work to shape their own design strategies. As new data and constraints emerged, the team continually reassessed feasibility and refined their approach, mirroring the realities of industry-style medical device development.

Throughout this process, this project has been supported by Professor of Integrated Engineering, Diana Chen, PhD, who has provided guidance through design reviews, technical discussions and project planning, supporting the growth and progress of the engineering team. 

Julia Marie Greely builds and tests the data acquisition circuit onsite at BD’s Research and Development lab

Working on the BD team has helped me experience incredible growth as an engineer." explains Greely. "It's strengthened my technical abilities, drawing on principles and lab methods I've learned throughout my classes; my communication skills, as my team and I are constantly in contact with each other and BD employees to ensure that we have what we need to hit every goal we set for ourselves; and has really given me a glimpse of what it's like to work like in the medical industry, specifically on the R&D side.

Greely continues, "I've learned so much from working interdisciplinarily, constantly running whiteboard sessions with the computer science team to make sure we're understanding each side of our project together. I'm really grateful to have such an amazing team and am excited to see how much we learn and where we end up after all of this." 

Jackson Downey (left), Phillip Banky (center) and Jennaya Horne (right) lead a whiteboard session to align hardware and software integration during early design planning

Looking ahead to the Spring 2026 semester, the biomedical engineering subteam will continue expanding the data collection and postprocessing workflow, with plans to scale data acquisition and further develop the artificial degradation process.

The computer science subteam will then train, test and evaluate the machine learning model as it is integrated with the infusion pump system. Through this next phase, the team looks forward to continuing to learn from one another and growing together as engineers while advancing the project toward a validated solution.

By Julia Marie Greely '26 (IntE)