HackDay Agra 2026

The Problem: AI often just "gives the answer," which kills critical thinking. We need tools that help students think, not just copy-paste. The Challenge: Create a study-buddy app that refuses to give the direct answer. Instead, it uses the Socratic Method. If a student asks, "How do I solve this physics problem?", the AI asks a guiding question back based on the student's previous mistakes. Key AI Use: Advanced system instruction (Role-play) + Long-term memory of student progress.

The Problem: Most online courses are linear. If a student is learning "React" but doesn't understand "Asynchronous JavaScript," they get stuck. Current platforms don't identify why a student is confused. The Challenge: A tool where a student uploads their syllabus and a specific topic they are struggling with. Gemini analyses the syllabus, identifies the "prerequisite gaps," and generates a custom, 15-minute "bridge lesson" to fix that specific gap.

The Problem: After orthopaedic surgery (like ACL or Knee replacement), patients often do their physiotherapy exercises wrong at home, leading to re-injury. Real-time human monitoring is too expensive. The Challenge: A web-app that uses the laptop camera to watch a patient perform exercises. By feeding video frames or coordinates to Gemini, the AI provides real-time voice feedback (e.g., "Lift your leg 10 degrees higher").personalised FAQ based on that specific report. Key AI Use: OCR + Medical Reasoning + Tone Shifting.

The Problem: After orthopaedic surgery (like ACL or Knee replacement), patients often do their physiotherapy exercises wrong at home, leading to re-injury. Real-time human monitoring is too expensive. The Challenge: A web-app that uses the laptop camera to watch a patient perform exercises. By feeding video frames or coordinates to Gemini, the AI provides real-time voice feedback (e.g., "Lift your leg 10 degrees higher").

The Problem: Designers often draw wireframes on whiteboards or paper. Moving these from a hand-drawn sketch to a working React/Flutter prototype takes hours of manual coding. The Challenge: A web tool where you take a photo of a hand-drawn UI sketch. Gemini Vision analyzes the layout, identifies buttons/inputs, and directly outputs the MERN stack code or Tailwind CSS to render that exact layout. Key AI Use: Vision-to-Code generation.

The Problem: Small startups struggle to maintain a consistent "voice" across Instagram, their Website, and LinkedIn. Their design doesn't match their story. The Challenge: A platform where a founder talks (voice-to-text) to the AI about their "Brand Vision." Gemini then generates a full Brand Bible: color palette suggestions (with hex codes), a 1-month content strategy, and specific prompts for Image Generators to ensure visual consistency.Key AI Use: Creative strategy + Cross-platform content orchestration.