Christoph Jurgen Shantz

Queen's MuLab Vertical Farm Data Dashboard Project Project Overview The Queen's MuLab Vertical Farm project is a collaboration between Christian Muise's MuLab and the Queen's Vertical Farming Team. For our Capstone Project our MuLab team was tasked with developing the software to store the data and streaming footage from the Queen's Vertical Farm from their Google Smarthome monitors and camera system. This project bridges the gap between agricultural data and researcher accessibility, enabling the QVFT to make data driven decisions in their operations. My key achievements on this project: I designed and implemented custom dashboards using Elasticsearch displaying critical agricultural metrics such as real-time precipitation monitoring, light, exposure tracking across growth zones, and temperature distribution visualization. These metrics set up the QVFT to easily correlate environmental factors with plant development and perform crop yield analysis. The result was a software that could make informed decisions for optimal growing conditions, intervene in plant growzones that were showing indicators of suboptimal yields, and forecast growth patterns and trends for crop yields. Data Management & Security I implemented the following security measures: Structured data storage protocols for the google home systems streaming data to logstash. I implemented secure deletion procedures with Role Based Access for the MuLake application as well as platform migration guidelines in the event that a later Capstone group would need to restructure the website from the ground up. These migration guidelines included Data backup and recovery protocols. Future-Proofing & Improvements Future proofed the MuLab project by proposing and documenting future enhancements such as data drift detection, enhanced security using Kerberos Authentication Tokens, automated anomaly detection using Neural Networks and TensorFlow, and adding RTOS using the native Kibana security features Technologies Used Frontend: Javascript and Kibana for visualization and user interface Backend: Elasticsearch, Logstash, Google home assistant Data Processing: Logstash for data pipeline management Dev Tools: Docker, GitHub

AnimeBible is an AI chatbot for anime discussions and recommendations, developed in collaboration with Queen's University alumni. The project combines Open-AI's GPT-4o model with finetuning methods to detect users intent and offer a wide range of services including: Anime recommendation based on users previous watches, plot summaries of recently released anime episodes, and solving powerscaling debates between different characters (Goku vs Naruto anyone?). Future additions to the platform include watchalong livestream series with Ishtar, the AI personality, and animation of anime fights using LumenAI. Some key technical achievements I've contributed to the project: Backend Development Implemented full backend deployment including Docker containerization Set up and managed PostgreSQL database for user information Integrated Redis caching for third-party cookies Deployed and managed website hosting on Google Firebase Developed custom Docker configurations with specialized flags for frontend/backend testing AI Integration & Optimization Currently working on fine-tuning and RAG (Retrieval-Augmented Generation) solutions Improved bot response latency and lookup speeds Enhanced accuracy for: TV show release date queries Plot synopsis retrieval Personalized show recommendations based on watch history Security & User Management Implemented security measures including Protection against SQL injection, Prompt injection, Cross-site scripting prevention and custom error handling Designed and implemented our secure login system using OAuth 2 Created user database architecture in Postgres, and configured Redis database for third party cookies from Anilist. Technical Stack Frontend: React.js Backend: Node.js, Express.js APIs: OpenAI API, AniList GraphQL API Databases: PostgreSQL, MongoDB Caching: Redis Hosting: Firebase Authentication: OAuth2 Development Tools: Docker, GitHub