Kaya AI India Hackathon 2026

Requirements

What to Build

Build an AI-powered solution for the construction industry in one of the four tracks:

🦺 Track 1: Physical AI

Construction sites are fast, dangerous, and overloaded with information, yet the people doing the work still operate with almost no real-time AI support.

Everything a worker needs to know lives in a document they can't open with their hands full on a beam. Meta smart glasses move the AI to where the work actually happens — they see what the worker sees and speak up when it matters.

Your challenge: build AI that can see the job site, understand what's happening in the moment, and help workers act faster, safer, and smarter.

We give you the hardware. What would you build if AI could see everything a worker sees?

🏗️ Track 2: Supply Chain

Materials get ordered. Then the chaos begins.

Teams are left asking the same questions on every project: What's been approved? What's being fabricated? What's delayed? Where is it now? Will it actually arrive when the project needs it? Nobody has good answers and delays cascade into missed schedules and blown budgets.

Your challenge: build an AI-powered system that brings real intelligence to the construction supply chain.

The data exists. The delays are real. The industry is waiting.

What would it look like if construction supply chains finally had a brain?

📋 Track 3: Procurement

Construction procurement runs on contracts, purchase orders, invoices, bids, and vendor relationships — and almost all of it still depends on manual review, disconnected systems, and people catching problems too late.

Your challenge: build an AI-powered system that makes construction procurement faster, smarter, and more resilient.

The complexity is real. The stakes are enormous. The opportunity is wide open.

What would procurement look like if every team had an expert analyst working 24/7?

🌍 Track 4: Open Innovation

Got an idea that doesn't fit the tracks above? Bring it.

Construction is one of the world's largest industries and one of its most underserved by software. The only bar: it solves a real construction problem, and it's something nobody has properly built yet.

Surprise us.

Go Deeper Than an API Call

The winning teams won't just wrap a chatbot around a prompt — they'll combine the right AI techniques with real problem-solving.

• Vision-Language Models (VLMs): Understand images, videos, drawings, and job-site conditions — not just detect objects.
• Document Intelligence: Extract insights from complex drawings, specifications, schedules, and contracts beyond basic OCR.
• Forecasting & Optimization: Predict delays, costs, demand, and risks using time-series, probabilistic, and optimization models.
• Agentic Systems: Build AI that can take actions — retrieve documents, send messages, make calls, and complete workflows.
• Construction-Specific AI: Use RAG, fine-tuning, and domain knowledge to make models understand your projects, vendors, SKUs, and specs.

Tip: Read recent research papers and explore open-source models — the biggest advantage often comes from using a technique nobody else thought to apply.

What to Submit (Stage 1 — by July 10)

 

• A written project proposal describing the problem, your proposed solution, and your chosen track
• A slide deck (Google Slides or PDF) — maximum 10 slides
• A 2-minute video walkthrough of your idea (unlisted YouTube or Loom link)
• Names and college affiliations of all team members (2–4 members per team)