The latest articles on DEV Community by Sanya Gupta (@bstok). https://dev.to/bstok https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4154636%2F2078eef6-6c7b-439b-b265-3643c0b2167e.jpg https://dev.to/bstok en Sanya Gupta Sun, 04 Oct 2026 13:07:15 +0000 https://dev.to/bstok/dropfolder-find-your-event-photos-from-massive-google-drives-instantly-31a1 https://dev.to/bstok/dropfolder-find-your-event-photos-from-massive-google-drives-instantly-31a1 <p>This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend</p> <h2> What I Built </h2> <p>Every time a college event, trip, or wedding wraps up, someone drops a massive Google Drive link containing thousands of unorganized photos. Finding your own photos usually means spending hours manually scrolling through dozens of subfolders.</p> <p>I built <strong>DropFolder</strong> to solve this exact problem for my friends. </p> <p>Instead of hunting for photos manually, you simply paste the public Google Drive folder link and upload a single selfie. <strong>DropFolder</strong> scans through all the subfolders in the Drive, runs open-source face recognition to match your face against every photo, and gives you a clean gallery of just your pictures ready to download.</p> <h2> Demo </h2> <p>The app is deployed separately across two services on Render:</p> <ul> <li> <strong>Frontend App:</strong> <a href="https://dropfolder-frontend.onrender.com" rel="noopener noreferrer">https://dropfolder-frontend.onrender.com</a> </li> <li> <strong>Backend API:</strong> <a href="https://dropfolder.onrender.com" rel="noopener noreferrer">https://dropfolder.onrender.com</a> </li> </ul> <h2> Code </h2> <p>Check out the full open-source codebase on GitHub:</p> <p><a href="https://github.com/BStok/DropFolder" rel="noopener noreferrer">https://github.com/BStok/DropFolder</a></p> <h2> How I Built It </h2> <p>The app decouples heavy multi-photo scanning into a stream-and-match pipeline:</p> <ol> <li> <strong>Google Drive Streamer:</strong> Parses the public Drive link, traverses all nested folders, and streams photo thumbnails directly in-memory using <code>PIL</code> and <code>io.BytesIO</code> without writing huge image files to disk.</li> <li> <strong>Open-Source Face Embedding:</strong> Uses <strong>InsightFace</strong> (ArcFace architecture on ONNX Runtime) to detect faces and extract 512-dimensional floating-point vectors for every detected face.</li> <li> <strong>NumPy Cosine Similarity Engine:</strong> Rather than relying on heavy vector database dependencies for an MVP, I represented face embeddings directly as <strong>NumPy matrices</strong>. When a user uploads a selfie, its target vector is compared against the event matrix using dot-product cosine distance math in milliseconds.</li> </ol> <h2> Why Does Open Innovation Matter? </h2> <p>Using open-source AI models and self-hosting over proprietary closed APIs (like AWS Rekognition or Azure AI Vision) was critical for three key reasons:</p> <ol> <li> <strong>Biometric Privacy &amp; Data Control:</strong> Facial embeddings and selfies contain sensitive biometric data. Proprietary APIs force you to send your friends' face data to corporate cloud servers. By running an open-weight model directly inside our self-hosted backend, photos and biometric embeddings remain under our control and are discarded after processing.</li> <li> <strong>Zero API Cost Escalation:</strong> Running closed APIs against event folders containing 2,000+ high-res photos incurs heavy per-image scanning and storage fees. Open-source models like InsightFace allow processing thousands of faces for $0 in API fees.</li> <li> <strong>Mathematical Freedom:</strong> Having direct access to raw 512-d feature vectors from open weights meant I didn't need black-box API lock-in; I could write a lightweight, hyper-fast NumPy matrix comparison script in just a few lines of Python.</li> </ol> devchallenge weekendchallenge hf26challenge