1Web3 Track 2 PS
Build decentralized applications (dApps), smart contracts, or blockchain-based solutions. Explore areas like DeFi, NFTs, DAOs, and decentralized identity while solving real-world problems using Web3 technologies.
Build a blockchain-based decentralized identity system that gives users full control over their personal data, enabling secure, privacy-preserving authentication across platforms without relying on centralized authorities.
Develop a decentralized autonomous organization (DAO) platform that allows communities to propose, vote, and govern decisions transparently using smart contracts.
2Business Requirements Document (BRD) Generation Agent 2 PS
Create an AI-powered agent that can automatically generate structured Business Requirements Documents from raw inputs such as client conversations, notes, or problem statements. Focus on clarity, automation, and real-world usability for businesses and product teams.
India is home to 1.4 billion people, hundreds of languages, and countless unsolved problems. Yet most AI tools are built for English speakers. Sarvam is changing that — and now it's your turn.
Use Sarvam's APIs to build anything that makes a real difference. A product, a tool, a service — the theme is completely open. The only requirement: use at least one Sarvam API and solve a problem that matters.
Build a scalable, multi-modal AI system using Google Gemini AI and integrated Google Cloud tools (such as Vertex AI, Cloud Storage, and BigQuery) that can process real-time, fragmented data (text, images, and documents) and deliver accurate, context-aware, and explainable decisions in complex and dynamic environments
3Intelligent Data Dictionary Agent 2 PS
Develop a smart system that can analyze datasets and automatically generate a comprehensive data dictionary. The solution should define data fields, types, relationships, and descriptions, making data more understandable and accessible for teams working with analytics and large datasets.
Build an Intelligent Data Dictionary Agent that automatically analyzes datasets to identify fields, data types, and relationships, and generates clear, human-readable descriptions to make data easy to understand and use.
Build an Intelligent Data Dictionary Agent that leverages AI to automatically interpret complex datasets and generate contextual metadata, including column definitions, data lineage, relationships, and usage insights. The system should enable natural language interaction, allowing users to query and understand datasets easily, while ensuring consistency, governance, and real-time updates across evolving data sources.
4Build a Real-Time Video AI Agent using TruGen AI 1 PS
Build a real-time video AI agent (using TruGen AI APIs) that users can call and interact with to complete a real-world task end-to-end by integrating external tools (via MCP/APIs), not just provide answers.
Build a real-time video AI agent (using TruGen AI APIs) that users can call and interact with to complete a real-world task end-to-end by integrating external tools (via MCP/APIs), not just provide answers.
Example Use Cases
-
Career Coach Agent
A real-time video agent that guides users on career decisions, answers questions, and provides personalized advice through live interaction. -
Code Reviewer / Tutor Agent
A real-time agent that helps students debug code, explain concepts, and teach interactively based on shared code or queries. -
Customer Support Agent
A real-time agent that resolves user issues via voice, screen sharing, or webcam by identifying problems and guiding solutions instantly. -
AI Interviewer Agent
A real-time agent that conducts mock interviews, asks questions, evaluates responses, and gives feedback live.
(These are just example use cases - you can build any creative real-time video AI agent that solves a real-world problem.
5Duality 1 PS
Duality AI is excited to present the Offroad Autonomy Segmentation challenge, a challenge designed to explore cutting-edge AI training techniques. Participants will train a model using annotated images of a desert environment, then test that model on a novel, but still desert, environment. All the data is generated from Duality AI’s digital twin simulation platform, Falcon, using FalconCloud’s geospatial-based digital twin environments. Each team will work to achieve the most accurate and precise model by adjusting training parameters and processes. Along the way, participants will learn how Duality AI uses industry-proven techniques and tools and high-quality synthetic data to train AI models for context shifts, unseen environments, or difficult-to-access data, such as remote areas. Through this competition, participants will develop novel AI training skills, enhance their portfolio, build connections with Duality AI and other participants, and have the opportunity to win prizes and recognition! Duality provides the problem statement ONLY. Hackathon organizers will handle all judging and prize allocation. Duality is not responsible for the event, judging, or prizes. Please contact the Hackathon organizers with concerns.
https://drive.google.com/file/d/1sYYGJM9n02j-6nYG6rFQOrfg7uJ7io32/view?usp=sharing
Registration Link: https://falcon.duality.ai/auth/sign-up?utm_source=hackathon&utm_medium=instructions&utm_campaign=AISummit
