iCloudEMS
Website:
icloudems.com
Job details:
Computer Vision Engineer — Smart Campus
iCloudEMS · Remote / Noida NCR · Full-time · Urgent
The problem you are solving
50,000 students. Hundreds of classrooms. Attendance marked on paper by a faculty member who has 90 seconds to call roll before lecture starts. It is inaccurate, gameable, and meaningless as data. We are replacing it with existing CCTV infrastructure — no new hardware, no wristbands, no apps. Just cameras that already exist, a model that recognises faces at scale in real-time, and a system smart enough to also tell you which rooms are overcrowded, which are empty, which have lights burning with nobody inside, and which need security attention. Campus intelligence, from cameras that are already there.
What you will build
Face recog.
Identify 50,000 enrolled students from CCTV feed — low-res, partial occlusion, varying angles
Attendance
Real-time auto-marking at scale, synced to the student ERP — no manual input
Crowd detect
Flag overcrowded rooms vs capacity threshold, real-time alerts to admin
Empty rooms
Detect booked-but-empty rooms and free them up automatically
Utilities
Lights on in empty room, fan running, AC on — flag for facilities team with camera snapshot
Edge + cloud
Low-latency processing at edge where needed, aggregation and analytics on cloud
Must have built something hard — any of these
Face recognition at 10,000+ identities
Real-time video analytics pipeline
YOLO / object detection production deploy
Multi-camera synchronisation system
Edge inference (Jetson / OpenVINO / ONNX)
Crowd / occupancy detection system
Competition — 3 rounds
Round 1
Given a set of low-res CCTV stills from a real campus: identify occupied vs empty rooms, count heads, detect lights-on-no-occupancy. Submit working code.
Round 2
48-hour sprint: extend your solution to handle a 10-camera simulated feed simultaneously. Latency under 2 seconds per frame is the bar.
Round 3
Live oral defence with Sachin and Hemant. Scale it to 500 cameras. Where does it break? How do you fix it?
Click on Apply to know more.