Adhar based Exam Attendance Verification
Budget / Salary₹75,000–150,000
TypeFreelance project
LocationRemote
Posted1 hour ago
I need a full solution that replaces paper attendance at semester exams with AI-driven face recognition. The job has two parts: a cross-platform mobile app for invigilators and a web dashboard for supervisors.
Mobile workflow
• Before each exam I will preload a spreadsheet (.csv or .xlsx) that contains student photos and details onto the device.
• At the venue the invigilator opens the app, points the phone camera at the student, and the app instantly compares the live face with the stored photo.
• If the confidence score is high, the student is automatically marked “Present” together with time-stamp and optional GPS tag.
• If the match is low the record is automatically set to “Supervisor Review,” blocking the student until a decision is made.
Web dashboard
Supervisors can upload the original spreadsheet, push exam rosters to devices, watch live attendance numbers, and review or override “Supervisor Review” cases from any browser. All data syncs once the mobile device regains connectivity.
Core requirements
• Native-like Android & iOS build (Flutter, React Native, or similar—your recommendation is welcome).
• On-device face recognition (TensorFlow Lite, FaceNet, or another efficient model) so the basic check works without internet.
• Secure REST or GraphQL API for syncing flagged records and attendance logs.
• Role-based access: admin, supervisor, invigilator.
• Exportable reports (CSV/PDF).
Deliverables
1. Source-controlled code for mobile app and dashboard
2. Compiled APK/IPA ready for store submission
3. Deployment scripts (Docker or equivalent) for the backend
4. Setup and user documentation
Acceptance criteria
• ≥95 % face match accuracy on the supplied photo set in normal classroom lighting
• Offline attendance capture for at least 4 hours without network
• Successful upload of a 2 000-student spreadsheet and real-time supervisor review of flagged cases
Let me know which tech stack you propose, any similar projects you have shipped, and an estimated timeframe for an MVP followed by full release.
Mobile workflow
• Before each exam I will preload a spreadsheet (.csv or .xlsx) that contains student photos and details onto the device.
• At the venue the invigilator opens the app, points the phone camera at the student, and the app instantly compares the live face with the stored photo.
• If the confidence score is high, the student is automatically marked “Present” together with time-stamp and optional GPS tag.
• If the match is low the record is automatically set to “Supervisor Review,” blocking the student until a decision is made.
Web dashboard
Supervisors can upload the original spreadsheet, push exam rosters to devices, watch live attendance numbers, and review or override “Supervisor Review” cases from any browser. All data syncs once the mobile device regains connectivity.
Core requirements
• Native-like Android & iOS build (Flutter, React Native, or similar—your recommendation is welcome).
• On-device face recognition (TensorFlow Lite, FaceNet, or another efficient model) so the basic check works without internet.
• Secure REST or GraphQL API for syncing flagged records and attendance logs.
• Role-based access: admin, supervisor, invigilator.
• Exportable reports (CSV/PDF).
Deliverables
1. Source-controlled code for mobile app and dashboard
2. Compiled APK/IPA ready for store submission
3. Deployment scripts (Docker or equivalent) for the backend
4. Setup and user documentation
Acceptance criteria
• ≥95 % face match accuracy on the supplied photo set in normal classroom lighting
• Offline attendance capture for at least 4 hours without network
• Successful upload of a 2 000-student spreadsheet and real-time supervisor review of flagged cases
Let me know which tech stack you propose, any similar projects you have shipped, and an estimated timeframe for an MVP followed by full release.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.