New disease outbreaks continue to occur annually, and health officials often struggle to analyze rapidly changing information across myriad sources and act fast. EpiNTelligence, by Summer 2026 Master of Information and Data Science alums Alex Kim, Alex Lei, Elaine Chan, and Sia Morris, is a tool that synthesizes fragmented data from existing sources such as the CDC, the WHO, and local health departments to accelerate decision-making by public health officials.
The team was awarded the Summer 2026 Hal R. Varian Capstone Award, which recognizes the semester’s top project.
To learn more, we interviewed the team —
What inspired your project? How did you decide on the concept?
Elaine: Going into the capstone semester, we were very excited about the opportunity to work on a socially impactful project in a domain that might not typically receive the most frontier technologies. The hantavirus outbreak was unfolding as our semester commenced, and it struck me how much inconsistency there was across government agencies, health reporting bodies, and the news.
Initially, we focused on signal gathering, but after interviews with local health professionals, we quickly realized the problem was not a lack of information but inefficiencies in quickly making decisions from vast volumes of unverified and often conflicting reports. That insight shaped our concept: rather than building another aggregator, we designed a system anchored on provenance, where every claim can be traced back to its source and assessed for trustworthiness.
Alex Kim: While the public health use case came from our team’s conversations with domain experts, what inspired me was the architecture behind the system. I found myself questioning whether some human-in-the-loop approaches preserve human judgment. If AI defines the options and the human’s role becomes reviewing or approving them, we’ve made people more efficient but not necessarily more thoughtful or creative. I wanted to explore a different approach: How can AI help keep smart people smart instead of outsourcing the thinking? EpiNTelligence was our attempt to answer that question.
“Build something weird. Build something that challenges your assumptions. If it keeps you curious, it’ll probably teach you more than you expected”
What was the timeline or process like from concept to final project?
Alex Lei: At the start, the biggest challenge was narrowing down our scope. We had so many potential directions, from maps for epidemiologists and AI agents for doctors, to rumor detectors and outbreak dashboards. While other teams jumped straight into exploratory data analysis (EDA) and modeling, we stepped back to get the core problem right. We cold-emailed subject matter experts (SMEs) across many platforms (local jurisdictions, UC Berkeley School of Public Health, LinkedIn, etc.) to understand how officers actually navigate outbreak reports. We didn’t even lock in our target audience until mid-semester, but those expert insights gave us the clarity we needed. Once we had that direction, everything clicked. We finalized our datasets, built the core architecture, and brought EpiNTelligence from concept to presentation-ready.
Elaine: Something we realized throughout this process is that health professionals will have conflicting opinions on how a process should be optimized or on accepting new technology. Before we shared our prototype with subject matter experts, it was often difficult for them to conceptualize how what we were building was distinct from what already exists, or whether it effectively solved a need in the workflow. However, once we showcased our prototype, our progress accelerated as we got more direct feedback on the usefulness of specific features relative to an epidemiologist’s actual research workflow.
How did you work as a team? How did you work together as members of an online degree program?
Elaine: Something that really worked for us was having each member own a different module of the deployment (i.e., model development, data pipeline, cloud infrastructure, UI and frontend). This let us asynchronously update and optimize our respective sections while minimizing bottlenecks in our team's overall workflow.
Alex Kim: Working in an online program required clear communication and trust, and I think that leadership rotation helped us build both.
How did your I School curriculum help prepare you for this project? Which course in the program was uniquely helpful in preparing you for your capstone project and why?
Elaine: In DATASCI 266, Natural Language Processing (NLP) with Deep Learning, we also chose our own final project topic, scoped more narrowly to our use of NLP, choice of models, and evaluation methods. The confidence and fluency we had built in model development let us quickly break EpiNTelligence down into small executable mini-projects and gave us significantly more creative freedom and flexibility with the customer-facing user experience and design of our end product.
Alex Lei: DATASCI 205 was uniquely valuable because it bridged the gap between local development and cloud execution. Learning to connect GitHub, Python, and the command line directly to Amazon Web Services (AWS) taught me how to manage real-world infrastructure, which is crucial when building our project’s data architecture.
Sia: Every course in the curriculum prepared us for our capstone project, and the variety of electives taken by our team helped us build EpiNTelligence. If I had to choose a course that helped me prepare most for capstone, I would pick DATASCI 221: Modern AI Strategy and Applications. DATASCI 221 demonstrated the importance of speaking with subject matter experts and thinking critically about AI use cases, which shaped how I approached our capstone.
Do you have any future plans for the project? Do you plan on continuing the work, and if so, what are you most excited about?
Elaine: I am super excited to continue working on this project. We received very encouraging feedback from our subject matter experts that has given us ideas for further features and parts of the workflow to target. The biggest challenge going forward may be working around information salience with more sensitive data, such as patient data or signal data owned by private organizations. Regardless, EpiNTelligence has given us a wonderful opportunity to rethink the existing public health information paradigm through the lens of what frontier technology can safely provide.
Alex Kim: I’ve been exploring this type of architecture for some time, breaking complex reasoning into smaller, more manageable processes that AI and humans can work through together. EpiNTelligence allowed me to test those ideas in a real-world setting. What excites me most is that the architecture is largely use-case-agnostic. While we demonstrated it in public health, I believe the same approach could support decision intelligence in many other domains where people need to reason through complex, uncertain, or conflicting information.
How could this project make an impact, or, who will it serve?
Alex Kim: I hope this project shows that AI can help people think better. Rather than optimizing humans for efficiency, I’d like to see AI preserve what makes us uniquely human: curiosity, creativity, and judgment. Although we built EpiNTelligence for public health, I believe this approach can benefit anyone solving complex problems with an open and curious mind.
Elaine: EpiNTelligence was designed not to offload critical thinking but to collapse the mundane, error-prone administrative steps around it. We hope this project can unload some of the burden that epidemiologists and researchers currently carry, and encourage them to take on more ambitious projects rather than be limited by the technology that currently exists. We also hope that by actively building the next generation of tools in this space safely and reliably, we can eventually shift the sentiment of distrust toward frontier technology in public health.
Sia: Many subject matter experts expressed mistrust in specific data sources and preferences for others. EpiNTelligence could impact clinicians by helping them see differences in what various government agencies and organizations report. Additionally, by automating data ingestion, EpiNTelligence provides a single repository of infectious disease knowledge, saving clinicians time elsewhere.
Now that you've finished your capstone, what advice would you give yourself at the start of the program? What do you wish Day One you knew?
Alex Kim: I wish I had known that it was okay to have fun with it. Some of the most meaningful learning came from building things simply because I couldn’t stop thinking about them. Build something weird. Build something that challenges your assumptions. If it keeps you curious, it’ll probably teach you more than you expected.
Elaine: Don't be afraid to start prototyping from Day 1! It's so much easier for subject matter experts to provide constructive criticism on something they can look at and play around with, even if not all of the backend is fully plugged in yet. That said, make sure to thoroughly research what is being built in your domain on a regular basis. It has never been cheaper and easier to build, and other projects in the space can provide signal and insight into pervasive gaps in the field, types of infrastructure to employ or avoid, and features that users have already validated or rejected.
Anything else you’d like to share?
Alex Kim: MIDS reminded me that learning doesn’t end with a degree. If anything, it made me even more excited about the questions I haven’t answered yet.
Elaine: Don't be afraid to build first and then iterate, refine, and fill in gaps in your knowledge as you go, understanding first principles more deeply along the way. At the end of the day, the limit of your potential is really the limit of your curiosity, so don't let an initial lack of deep understanding of how something works hold you back.
Alex Lei: Never stop learning!
Sia: We would like to thank our instructors, Joyce Shen and Korin Reid, for their guidance and support. We also thank all the subject matter experts who gave us valuable feedback and helped shape EpiNTelligence. Go Bears!
