Today’s job searchers know firsthand how important tracking job applications is when hundreds of applications are involved. However, keeping track of interview dates, resume versions, and other helpful metadata wastes time that could be better spent on completing more applications. Using a spreadsheet or another rote mechanism for inputting each application is common but far from an ideal solution. Willow reduces the labor involved in tracking applications by offering an email-based forwarding system to organize and analyze job applications, enabling users to spend more time on applications not admin.

Harmony Hub is a web-based application designed to alleviate tinnitus symptoms during music listening by applying split-ear notch filters at user-specified tinnitus frequencies. A key challenge was balancing audio quality with filtering to ensure a satisfying listening experience. We leveraged Creative Commons music to support underground artists and maintain compliance with commercial licensing standards. The outcome is a personalized platform that reduces the perceived intensity of tinnitus while promoting emerging musicians. Harmony Hub empowers users to manage their symptoms and increase emotional music connectivity, offering a unique intersection of auditory therapy, user-driven customization, and ethical music streaming.

Slate addresses the core challenges faced by indie filmmakers—fragmented communication, constant rescheduling, and costly production tools. By automating scheduling, parsing scripts into scenes, and centralizing crew communication, Slate streamlines pre-production workflows. The outcome is a significant reduction in time spent coordinating logistics and fewer missed updates. Slate empowers filmmakers to focus on storytelling, not scheduling, making small productions more efficient and less stressful.

Maintaining security patches across Android branches requires constant developer effort to manually adapt patches and resolve complex merge conflicts. With hundreds of security fixes incoming each month, this process is slow, error-prone, and diverts valuable engineering time from higher-impact work. Vidar transforms patch maintenance with AI-powered automation, using large language models to resolve rejected patches and generate clean, context-aware fixes. By eliminating much of the tedious manual intervention, Vidar accelerates security updates, reduces developer workload, and ensures vulnerabilities are patched quickly and reliably across diverse downstream branches, strengthening the security of the Android ecosystem without disrupting ongoing development efforts.

The Point-in-Time (PIT) Count is a count of sheltered and unsheltered people experiencing homelessness in a location on a single night. Surveys are often used by governments and research organizations to understand the needs of unhoused populations and are critical for the PIT count. However, current existing versions of the PIT count are labor-intensive and often undercount populations in inaccessible locations. In 2023, our sponsors developed a new data collection method for the count that utilizes Respondent Driven Sampling, leveraging unhoused individuals’ social networks to collect data and estimate population size/demographics. The deliverable of this capstone project is a customizable, open-source, and easy-to-use survey app that meets the needs of all stakeholders of the PIT Count. Our implementation enables scanning referral QR codes, generating and completing new surveys, and printing linked referral codes using tablets and portable printers.

Abstract: RaiseYourYAR is a standardized, automated tool built to streamline faculty workload reporting for the Department of Urban Design & Planning at UW. It replaces manual data entry and collection with a centralized system that gathers, organizes, and visualizes faculty activity data. By eliminating manual work and repetitive tasks, YAR saves time for faculty while equipping department chairs with actionable insights for strategic planning and decision-making. The outcome is a scalable, long-lasting solution that improves operational efficiency and supports data-driven leadership.

Less technical employees in Pierce County often struggle to accurately interpret data-driven reports, leading to miscommunication and gaps between technical and non-technical teams. To address this, we designed an accessible, beginner-friendly data literacy resource tailored to staff with little to no technical background. Rather than functioning as a structured course, this modular guide is designed to serve as a quick reference tool for employees. The outcome is a learning platform that breaks down key data concepts using real-world and workplace-relevant examples. This initiative helps bridge communication gaps, assists employees to engage with data more confidently, and supports more effective collaboration across county departments.

Social media has created challenges for banks who are addressing reputational risks. To address these concerns, banks require more advanced tools and techniques. We built a solution using AI/LLM to conduct real-time sentiment analysis for financial institutions. This solution provides a real-time data analysis and visualization dashboard that allows the organization to understand and respond to the problems and outages that they are facing. This project reduces management costs while increasing the speed and accuracy of incident responses by automatically generating a suggested mitigation plan, allowing for stronger coverage of reputational risk.

This sponsored project by Microsoft supports the Customer Support and Services (CSS) Team, which helps customers resolve technical issues and maximize the value of Microsoft products. Z.AI, the team’s internal AI tool, initially faced challenges due to its steep learning curve and usability limitations, making prompt engineering difficult for many users. To address this issue, we redesigned Z.AI with a more intuitive interface and improved user workflows. This allows the CSS team to more easily create, test, and iterate on AI prompts, enhancing their ability to support customers efficiently and effectively.

In collaboration with Microsoft’s Digital Operations Experience Team, we designed the first desktop and mobile dashboards for Azure IoT Operations that incorporated Copilot AI to support Operations Technician users. As production lines may have thousands of machines across different sites, our goal was to simplify data monitoring and reduce production downtime by providing AI-generated summaries that prioritized actionable insights, predictive forecasts, and root-cause explanations to keep users informed. Our work lays the foundation for scalable, accessible, and AI-driven operational tools in this production line space.