Pajarito Environmental Education Center (PEEC)

first time in a long time i am excited to share something. watch the video with volume turned to mid ... its damn fun xD ... just do it! so we hacked this time at Opportunity Hack, and i seriously just love them. Last year attended my very first hackathon with them, and m just blown away with their meticulous attention to detail. Its like a smooth ride of enjoyment of 24-36 hours I just love them. this year we built a custom search engine for Casey Lundberg's gigantic collection of photos in Dropbox. The NGO she works for is Pajarito Environmental Education Center (PEEC). and this was the problem statement me and Rajat Pawar were looking for. A >100GB live dataset to create some sort of index/fabric on top of for fun hack. And that's exactly what happened. So Casey, had photos in her Dropbox, but no way to search them based on content. And no way to generate automated tags for searching them either. So we built one for her. We get the image, generate tags, caption, title (gpt-4o-mini), extract metadata, generate embeddings (local) , store all of this on local PG and delete the image from our machine, because it is stored in her Dropbox. Later when she searches, we just search through the index to get the dropbox url, and show the results to her with thumbnails. - Local PG, because we could just recreate the db, if something happens because the photos are still safe in dropbox. This was all possible by first and foremost my team: - Data Engineer Master: Rajat Pawar - Machine Learning and Data Analysis Expert: Saurabh Zinjad - Our Full Stack Developer: Harsh Victor Challa the organizing team: Greg Vannoni and mentors Opportunity Hack, Serina Caballero, Mark and Radhika Bhati for their mentorship. And ofcourse Casey, for keeping us on track and guiding us perfectly through her problem And now for the ultimate entertainment watch the video below, and then view our website demo at: https://lnkd.in/gK3-m4HX --- but srsly, just watch and enjoy the video audio credits: https://lnkd.in/gwtXRvrn

🚀 Excited to share that our team was placed in the top 5 at Opportunity Hack 2024! 🚀 Our project was built for the Pajarito Environmental Education Center (PEEC), a nonprofit dedicated to enriching lives through environmental education at the Los Alamos Nature Center in New Mexico. PEEC faced significant challenges in managing their growing media library, relying on time-consuming manual processes without automated tagging or backup solutions. To address this, we developed a cloud-based media management system with AI-driven tagging, content-based search, and seamless integration with existing tools. 🔍 What We Built: Our system automatically tags, organizes, and retrieves media files through AI models like CLIP and 4o-mini, significantly reducing manual effort. It enables users to upload media from sources like Dropbox, and provides a robust backup solution using AWS. The intuitive React.js frontend makes it simple for non-technical users, while the FastAPI and Postgres backend ensures efficient processing of search queries. 🛠️ Tech Stack: Frontend: React.js (JavaScript) Backend: FastAPI (Python) Cloud Infrastructure: AWS Database: PostgreSQL AI Models: CLIP, 4o-mini  OAuth and API Integration: Dropbox 🌟 Key Accomplishments: -92% reduction in manual labor for media organization, freeing up valuable resources for PEEC to focus on educational initiatives. -Seamless AI-driven media tagging and search capabilities that allow PEEC to retrieve images based on descriptions like “people outside” or “child with binoculars.” -Integrated Dropbox API for easy media uploads and sharing, ensuring files are backed up and easy to access across platforms. -Developed a system that is scalable and future-proof, with open APIs to integrate with PEEC's growing suite of tools. Our solution allowed us to apply our expertise in AI, cloud services, full-stack development, and API integration, showcasing our ability to deliver cutting-edge, fully deployed solutions that are ready for real-world use. We couldn’t have achieved this without the incredible support of our mentors — Greg Vannoni, Radhika Bhati, and Mark O. — as well as the representatives from PEEC, including Casey Lundberg and Serina Caballero, who provided invaluable insights and feedback throughout the hackathon. A heartfelt thanks also goes out to the many others who guided us along the way. Meet my teammates: Rohan Awhad, Saurabh Zinjad, and Rajat Pawar. We are incredibly grateful to the sponsors who supported Opportunity Hack , including ReadMe, MX, Meta, Spotify, and Devils Invent at Arizona State University. Their contributions made this event possible and allowed us to bring our project to life. Check out our project: 🔗 Devpost - https://lnkd.in/gbyjdWAf 🔗 GitHub Repository - https://lnkd.in/g5qxvxFW