7903 research outputs found
Sort by
Once Upon a Time for the Soul: A Review of the Effects of Storytelling in Spiritual Traditions
Get PDFBIOME to Transform Cooking
No full textIndividuals with dietary restrictions or specific health goals often struggle to find recipes that meet their needs. This challenge is further compounded by the lack of time and convenience many busy individuals face when planning and preparing meals. The BIOME app addresses this issue by o↵ering a mobile application that delivers personalized recipe recommendations based on user preferences, dietary restrictions, and ingredient availability. It also allows users to order ingredients directly through Instacart.
BIOME features a user-friendly frontend built with Flutter and a backend powered by Google Firebase, integrated with unsupervised machine learning to provide tailored content. A clustering algorithm was used to group similar recipes based on ingredients and nutritional information. The model’s e↵ectiveness was evaluated using metrics such as the silhouette coefficient and Davies-Bouldin index to ensure distinct and meaningful groupings of recipes. BIOME connects to the Instacart API to enhance usability, allowing users to convert ingredient lists into shopping links for faster, easier grocery access. Internal testing with mock users and a beta testing phase with friends and family helped identify bugs and gather feedback to improve performance and user experience.
Overall, BIOME demonstrates how machine learning and mobile design can support healthier eating habits by offering better recipe recommendations and reducing the time and effort needed to cook nutritious meals
BroncoNest
No full textSelecting suitable housing is a critical yet often overwhelming aspect of the college experience, particularly for incoming students who lack access to transparent and meaningful information about residential life. Existing platforms, such as official university housing portals or external real estate websites, fail to capture the nuanced, day-to-day experiences that influence student well-being, including cleanliness, social atmosphere, and sense of community. This issue is further compounded by the growing prevalence of remote decision-making and scattered, unverified sources of housing feedback.
To address this challenge, we developed BroncoNest, a scalable, cross-platform mobile application designed to centralize and personalize the student housing search process. Built using Flutter and Firebase, the platform integrates both on-campus and off-campus housing data, incorporating verified student reviews and intelligent filtering mechanisms. At its core, BroncoNest employs a Retrieval-Augmented Generation (RAG) model, hosted on a Flask-based Python backend and deployed via Firebase Cloud Functions, to deliver AI-driven housing recommendations tailored to user preferences.
The system supports role-based access control (RBAC), ensuring content authenticity and moderation through institutional oversight. Students authenticate using their university-issued email addresses, enabling them to contribute structured reviews, engage in real-time chat with peers, and save personalized housing selections. Additionally, university administrators are provided tools to manage listings and monitor platform activity. The backend architecture leverages Firestore for real-time database operations and Pinecone for vector-based similarity search, allowing the AI assistant to rank listings according to review-derived attributes.
By combining robust data infrastructure with machine learning and user-centered design, BroncoNest enhances transparency, fosters student connection, and supports informed housing decisions within the university ecosystem
Documentation-Aware Code Generation Via Retrieval Augmented Generation
No full textModern large language models (LLMs) increase the speed of software development but frequently hallucinate code, forcing developers to debug and rewrite generated code manually. This project introduces Documentation-Aware Code Generation, a retrieval-augmented generation (RAG) framework that injects authoritative documentation into the LLM prompt to reduce hallucinations in generated code. We build an ingestion pipeline that scrapes, cleans, chunks, and embeds official API documentation into a Pinecone vector database, then design a retrieval and re-ranking pipeline that retrieves the most relevant snippets for each user query. Results show that augmenting prompts with documentation lowers code hallucination by up to 60%. The system is exposed through backend APIs and a Next.js frontend, offering developers a tool that reliably generates code
High-Temperature Inert-Environment Thermomechanical Testing Furnace
No full textHigh-temperature materials testing is critical for frontier technologies, especially in the aerospace and power generation sectors. The objective of this senior design project is to research and develop an inert-environment sealed furnace for a lab mechanical tester to reach 2000°C for the Santa Clara University Materials Science Lab. Two groups of project members were created to address engineering design and manufacturing issues in various subsystems of the final vision: Team A focuses on structural analysis and fabrication, instrumentation, and electrical design, while Team B focuses on thermal, busbar, and fluid cooling design. Financial and logistical aspects of the project have been taken into consideration, and the design, analysis and verification of various components of the system without direct testing of a final product will inform future teams on the next steps
