two rounds 30 min back to back, with the team AeroDocs GitHub Link Technologies used: Next.js, TypeScript, Python, Node.js, Machine Learning, Ollama LLaMA 3.2, YOLOv8 • Built a secure knowledge base and VPC solution for Honeywell using a RAG-based vector search pipeline. • Implemented document search via barcode, part numbers, and image recognition using LLaVA and YOLOv8. • Enabled upload and retrieval of technical documents to streamline repair workflows. Object Detecting Vest GitHub Link Technologies used: Python, Raspberry Pi • Engineered an innovative object-detecting vest aimed at enhancing safety and navigation for the visually impaired. • Developed a mobile application for step tracking and seamless control of vest features. • Integrated hardware and software for optimal functionality and user experience. Schr¨odinger’s AI GitHub Link Technologies used: Next.js, TypeScript, Node.js, MongoDB, AWS, Vercel, Docker • Built a hallucination-resistant AI chatbot using RAG and factual verification layers. • Integrated LLM APIs with custom prompt engineering and a reliable knowledge base. • Reduced hallucinations by 40% and deployed a scalable backend on AWS and frontend on Vercel. Orbit GitHub Link Technologies used: Python, React, TypeScript, AWS, MongoDB, Vercel, Docker • Built a chatbot for real-time travel planning with live flight and booking data. • Used Gemini 1.5 Flash and FlightLabs APIs; improved response time by 50% and user satisfaction by 25%. • Backend servers currently hosted via AWS and frontend hosted via Vercel