Present-Sir

🎯 Inspiration Attendance — one of the most basic administrative tasks in any institution — is still broken everywhere. Teachers waste 10 minutes every class calling names. Students game QR codes by sharing them with friends. Proxies run rampant. Admins drown in spreadsheets. We asked ourselves: what if attendance could be invisible, fraud-proof, and actually intelligent? That question became PresentSir.

🚀 What it does PresentSir is a universal AI-powered presence intelligence platform that works across schools, universities, companies, and events — all from one web app.

Smart QR Check-in — rotating QR codes that expire every 20 seconds, tied to specific sessions Face Recognition — camera-based face verification at check-in, showing student name, roll number, and match percentage live on screen Active Presence Tracking — detects if a student switches tabs or minimizes the app and logs their real time-in-class in minutes AI Insights powered by Groq — teachers and admins get plain-English AI generated summaries of attendance patterns, at-risk students, and actionable recommendations Sentiment Pulse — students tap how they're feeling during class; teachers see a live mood summary Excuse Submission Flow — students submit excuses, teachers approve or reject with one click Attendance Calculator — students know exactly how many classes they can skip and stay safe Campus Events Board — admins post events, students read and know who to contact Exam Notice Board — timetables and exam notices posted by admin, visible to all students Multi-org support — School, University, and Company modes each with their own role-based dashboards Export to CSV — download any session's attendance records instantly

🔨 How we built it We started by researching real pain points from teachers and professors on Reddit and built our feature list entirely around what educators actually complain about — not what looks good in a demo. We designed the full UI first using v0 by Vercel, then exported the codebase to VS Code and wired up the real backend using GitHub Copilot for implementation with Claude guiding the architecture and prompts throughout. The face recognition runs entirely in the browser using face-api.js with no server-side ML needed. AI insights are generated on demand via the Groq API using Llama 3.

😤 Challenges we ran into

QR fraud prevention — making QR tokens truly single-use and session-bound without a native mobile app was tricky face-api.js in Next.js — getting browser-based face detection to work in a server-side rendered Next.js environment required careful client-side loading and Suspense boundaries Role-based routing — managing three org types, each with multiple roles (admin, teacher, student, manager, employee) and their own dashboards without routing conflicts was complex Build errors at deployment — useSearchParams() Suspense boundary issues in Next.js 16 caught us at the last minute during Vercel deployment Time — we built this entire platform in under 48 hours

🏆 Accomplishments we are proud of

Face recognition that runs 100% in the browser with live bounding box, name, roll number, and match percentage — no backend ML server needed A single web app that genuinely serves schools, universities, and companies with completely different workflows under one roof AI insights that turn raw attendance numbers into real plain-English advice for teachers and admins using Groq's Llama 3 The active presence tracker — knowing not just who showed up but how long they actually stayed is something no existing attendance system does

📚 What we learned

Browser APIs like Page Visibility API and getUserMedia are incredibly powerful and underused in edtech Prompt engineering for GitHub Copilot is a skill — vague prompts produce broken code, detailed prompts produce working features For a hackathon, shipping a polished demo of 70% of features beats shipping a buggy version of 100% Groq's Llama 3 is genuinely fast enough for real-time AI features in a web app

🔮 What's next for PresentSir

Real-time WebSocket presence — live heartbeat dashboard showing exactly who is online second by second GPS geofencing — confirm students are physically inside the classroom, not just on the same WiFi Device fingerprint lock — prevent the same QR from being used on multiple devices simultaneously LMS integration — sync attendance directly with Canvas, Moodle, and Google Classroom Mobile app — native iOS and Android for faster check-ins and push notifications Predictive analytics — ML model that flags students at dropout risk weeks before it happens Blockchain attendance records — tamper-proof, verifiable attendance certificates for exams and certifications