Noah was born and raised in Aiea, Hawaiâi. He graduated from Kamehameha Schools Kapalama and is currently a sophomore at UH Manoa majoring in computer science. He plans to pursue a career in software engineering and hopes to help people with his skills. In his free time, Noah enjoys programming, cooking, going to the gym, enjoying nature, and watching good TV shows and movies.
Home Island: Oâahu
High School: Kamehameha Schools Kapalama
Institution when accepted: UH Manoa
Project title: Building a Project-Knowledge Management System Using Large-Language Model Agents for the Gemini ObservatoryÂ
Project Site: Gemini Observatory, Hilo, HI
Mentors: Patrick Parks & Winston Wu
Project Abstract:
Gemini Observatoryâs technical staff face persistent challenges in maintaining organized and accessible project information. Essential data is currently distributed across multiple platforms such as Grafana, the Gemini Engineering Archives, and Channel Access, making it time-consuming and tedious for staff to locate and synthesize critical information. Recent advancements in large language models (LLMs) present a promising opportunity to address this issue. LLMs excel at processing and interpreting vast amounts of unstructured text, making them well-suited for knowledge management tasks. However, LLMs can only generate text based on the context directly included in their prompt, which limits their ability to access external systems or execute actions. This project explored the development of a centralized platform that integrates LLMs with Anthropicâs Model Context Protocol (MCP), enabling real-time interaction with existing observatory tools and data sources. This transforms the LLM from a passive text generator into an active assistant capable of intelligently selecting and using tools to fulfill user requests. The platform was designed to be modular and extensible so that new tools, data sources, and capabilities can be added or removed as project needs evolve. By combining the reasoning capabilities of LLMs with MCPâs ability to execute external commands and retrieve live data, this system could significantly reduce the time and effort required to locate, interpret, and act on observatory data. Ultimately, it will enable technical staff to interact with complex systems and data through natural language, improving efficiency, accuracy, and user experience.