Pre-AI Summit Hackathon

About the Hackathon: This is a 24-hour pan-India hackathon conducted as one of the Pre–AI Summit Activities at IIT Ropar, ahead of the India AI Impact Summit 2026 The hackathon is designed to expose student teams to real-world AI problems and accelerate the development of prototype solutions under mentorship from faculty and domain experts Participants will work on selected AI-for-social-good themes and build a functional solution where a prototype or MVP is acceptable All submissions must include reproducible code and benchmarking Themes: AI for agriculture AI for education AI for low-light object detection Note: Detailed problem statements, datasets, evaluation protocol, and input/output formats will be revealed on-site at IIT Ropar This information will be shared 1 hour before the hackathon begins to ensure fairness Eligibility: Open to undergraduate, postgraduate, and PhD students from India Team Size: 1 to 4 members per team Timing & Venue: Reporting time: Registration and inauguration at 9:30 AM on February 6 Campus exploration and lunch to follow Hackathon duration: Starts at 3:00 PM on February 6 Ends at 5:00 PM on February 7 Attendance requirement: At least one team member must attend registration and inauguration All team members must report by 3:00 PM Venue: School of AIDE, IIT Ropar Compute Support: Participants must bring their own laptops Teams should be prepared for quick environment setup Access to high-end cloud GPUs will be provided as per availability GPU access will be through participants’ laptops Prizes & Certificates: Total prize pool: ₹1,05,000 Winner: ₹25,000 Runner-up: ₹10,000 Certificates will be provided to every participant What You Build & Submit: A working solution where prototype or MVP is acceptable A public GitHub repository containing code and documentation Results or output files in the format shared at problem release Optional demo or application, if applicable to track bonus Submission Rules: Registration on Unstop is mandatory before the competition starts All development must begin only after the hackathon officially starts Pre-written code is not allowed, except for: Open-source libraries Framework boilerplates Public APIs There will be 5 checkpoint evaluations conducted every 4 hours, followed by a final submission A maximum of 2 final submissions is allowed per team across all tracks Any plagiarism or copying of existing projects will lead to immediate disqualification Judges’ decisions will be final and binding Special Note for Theme: AI for Education: Participants are expected to explore rPPG extraction methods before the hackathon During the hackathon, teams may use any 3 out of the following 7 rPPG algorithms: CHROM POS DeepPhys PhysNet EfficientPhys PhysFormer TS-CAN Exact task definition and evaluation details will be shared on-site Additional Information: More detailed guidelines are available in the official rulebook shared with participants Best wishes to all participating teams