The Merritt Digital Preservation System consists of 7 microservices running across 26 servers running in AWS. This system has been in operation for over a dozen years. For the past 5 years, our team has been on a DevOps journey. Our goal is to deploy all of our services using infrastructure as code. Once complete, we want to be able to scale each of our microservices up and down to meet current processing demands. Our team consists of 3 developers and a product manager.

As research regulations grow more complex, ensuring high-quality IRB (Institutional Review Board) submissions is more critical—and more time-consuming—than ever. In this session, learn how UCSF is partnering with Huron to co-develop an AI wrapper that enhances the Huron IRB system by validating determinations, automating checklists, and flagging compliance gaps for FDA-regulated research, with the ultimate goal of reviewing the entire application prior to its submission. But this session isn’t just for IRB professionals—anyone working with complex systems can learn from our approach.

Severe aortic stenosis (AS) is a life-threatening condition with a high mortality rate if left untreated. Studies show that symptomatic severe AS can have a 50% mortality rate within one year, yet timely intervention with aortic valve replacement (AVR) significantly improves survival outcomes. Despite the availability of effective treatment, many cases go unnoticed or unaddressed due to gaps in clinical workflows and provider follow-up.

This poster will highlight the impact of the APeX Enabled Research (AER) team on implementing clinical and technical innovations at UCSF. The poster will explain the primary goals of the team, volume of project work, and breadth of technologies used to support research innovation. Ideally UC Tech attendees will learn more about what is technically possible today to support research, get inspiration for projects at their organization, get connected with others doing similar work, and lead to future collaboration and innovation across UC campuses.

In the world of hybrid and remote work, collaboration can sometimes feel like trying to perform synchronized swimming, while miles apart. In this talk, we’ll dive into the art of building trust, fostering knowledge-sharing, and aligning teams in a digital environment. Just like synchronized swimmers, remote teams rely on coordination, communication, and routine to stay afloat. We’ll explore how to keep your team performing great stunts, even when working from different time zones or behind different screens.

This proof of concept integrates Large Language Models (LLMs) with Oracle PeopleSoft Campus Solutions to enable natural language querying of class and course data, democratizing access to data. PeopleSoft manages essential student records and supports critical academic processes. This prototype, built with Python, LangChain, the ChatGPT API, Streamlit, prompt engineering, and schema transformation, offers a chatbot-style interface for users who may not be SQL experts to access data.

Inaccessible digital content is a huge problem, but how do we describe and measure the problem well enough to formulate a strategy to remediate it? As we work toward the 2026 deadline to comply with federal digital accessibility rules, we're all feeling like the blind men in an Indian parable trying to describe an elephant for the first time: https://en.wikipedia.org/wiki/Blind_men_and_an_elephant. They describe the part they touch as if it represents the whole animal. Arguing, they can't reach consensus.

At UCLA, academic technology teams partnered with the Teaching and Learning Center to analyze instructor user stories on media use within the Canvas LMS. With awarded OpenAI licenses, the teams leveraged AI to organize and code interview data, uncovering key insights into how media enhances lectures, assessments, student projects, and active learning. Through strategic prompt engineering, raw interviews were transformed into actionable themes.

As institutions continue to evolve in the digital landscape, optimizing IT support is crucial for enhancing user experience and operational efficiency. This session explores how an AI-powered chatbot can revolutionize Epic user support by providing immediate assistance, reducing helpdesk call volume, and improving response times. A real-time dashboard tracks key metrics such as response time, interaction frequency, and peak call volume, enabling data-driven decision-making to refine support strategies.

Ensuring the safety of students and staff during emergencies is crucial. Our innovative platform leverages data analytics to enhance emergency preparedness in educational institutions. By integrating data from multiple sources, we create a golden contact record for each individual, ensuring reliable contact information. Utilizing GIS mapping and geocoding, it provides precise location tracking and targeted responses. Our automated notification system streamlines communication, reducing response times and improving efficiency.