Modeling Core Codeathon Showcases AI-Driven Data Exploration

The 2026 Modeling Core Codeathon brought together researchers, clinicians, data scientists, graduate students and engineers for a collaborative sprint focused on transforming how investigators interact with complex biomedical data. Held over three sessions spanning late March through mid-April, the event challenged teams to build tools that make multimodal research datasets more accessible, intuitive, and explorable through natural language interfaces and interactive visualization.

The initiative leveraged multimodal data resources from the SCRIPT U19 Systems Biology Center, including structured electronic health record data, clinical notes, imaging reports, single-cell datasets, Karius sequencing data and other integrated research modalities.

Participating teams delivered concise three-minute demos followed by hands-on exploration sessions in which judges tested functionality, usability and innovation. Projects were evaluated by an interdisciplinary panel of judges including Sasha Misharin, MD, PhD, Associate Professor, Medicine (Pulmonary and Critical Care); Scott Budinger, MD, Chief of Pulmonary and Critical Care in the Department of Medicine; Rich Wunderink, MD, Professor, Medicine (Pulmonary and Critical Care); Ankit Agrawal, PhD, Research Professor, Department of Electrical and Computer Engineering; and Nandita Nadig, MD, Associate Professor, Medicine (Pulmonary and Critical Care). Judging criteria emphasized usability, creativity, scientific impact and the ability to support real-world clinical and translational research workflows.

Project Highlights

Teams tackled a diverse set of challenges across clinical data exploration, AI agents, visualization and multimodal integration. Among the featured projects:

• Agent-Based Exploration Tools showcasing autonomous workflows for navigating and summarizing research datasets -- Sayak Chakrabarty
• Clinical Dashboard detailed patient-level drilldowns and many slices of layers of analysis – Wan-Ting Liao, Saki Amagai, Amy Krefman, Diane Lee
• Interactive Data Browser and Paper Integration focused on connecting datasets with relevant scientific literature and enabling intuitive discovery workflows – Catherine Gao
• Slack-Integrated Single Cell Data Assistant designed to provide real-time updates and interaction around single-cell research datasets directly within collaborative communication workflows – Max Schleck

Looking Ahead
The event concluded with prize announcements, celebratory recognition, and a collaborative discussion on next steps and future development opportunities. Organizers emphasized that many of the prototypes developed during the Codeathon demonstrated real promise for continued development and broader deployment within research workflows.

Participants 
Catherine Gao, MD, MS, Associate Director, Center for Collaborative AI in Healthcare
Luisa Cusick, Research Data Analyst, Division of Pulmonary and Critical Care Medicine
Alexander V Misharin, MD, PhD, Associate Professor, Medicine (Pulmonary and Critical Care)
Amy Krefman, PhD student in Health and Biomedical Informatics (HBMI)
Courtney Reamer, MD, Department of Pathology Fellow
Max Schleck, Research Data Analyst Associate, Division of Pulmonary and Critical Care Medicine
Karolina Senkow, PhD Student
Saki Amagai, PhD Student
Wan-Ting Liao, Research Data Analyst
Diane Lee, PhD Student in Health Informatics
Vijeeth Guggilla, MD-PhD Student
Sayak Chakrabarty, PhD candidate, Department of Computer Science