The Golden Hour - A Voice AI Hackathon

Hosted by: Entrepreneurship Cell, IIT Kharagpur in Association with CallKaro AI TechForge is an applied artificial intelligence competition that immerses participants in the rapidly evolving landscape of enterprise-grade Generative AI and Conversational Voice systems. The event challenges participants to design, build, and deploy intelligent voice agents capable of handling real-world enterprise communication and customer interactions under operational constraints. Participants will explore LLM architectures, prompt engineering, speech-to-text/text-to-speech pipelines, and conversation flows, engineering agents that operate reliably within strict business guardrails. Designed for AI enthusiasts, developers, and product builders, TechForge tests system design, algorithmic rigour, prompt engineering, and the practical execution of autonomous voice agents. Prize Pool: INR 1,00,000 + PPI Opportunities Winner: ₹50,000 + Certificate of Excellence Runner Up: ₹30,000 + Certificate of Excellence Second Runner Up: ₹20,000 + Certificate of Excellence Top 10 submissions: Potential PPI Opportunities at Callkaro.AI Eligibility Criteria: Teams can consist of 1 - 6 members. College/Inter-College students can participate. Students from any program or year of study may register. No changes in team composition will be allowed after registration. Competition Structure: Round 1: AI, Generative AI & Conversational Voice AI Quiz (Online): An online, beginner-friendly assessment covering core principles of Artificial Intelligence, Generative AI, Large Language Models (LLMs), prompt mechanics, context management, hallucination mitigation, Voice AI pipelines (STT/TTS), conversational design, and enterprise calling workflows. Maintains an even balance between foundational LLM concepts and real-world conversational voice agent architectures. Shortlisted participants/teams advance to the enablement and build stages. Hands-on Masterclass: Building AI Calling Agents with CallKaro AI (Online): A dedicated technical workshop conducted by industry experts from CallKaro AI. Covers agent architecture, prompt design patterns, state machine conversation flows, dynamic variable configuration, real-time testing, and platform access/credits for the development phase. Round 2: Agent Development & Failure Mode Testing (Online): Teams receive curated industry problem statements to design and deploy a functional voice calling agent. The agent must operate strictly within predefined guardrails regarding pricing disclosures, service availability, and explicit customer consent. Teams must test their agent against at least five distinct simulated-customer scenarios of their own design, including at least one edge-case scenario engineered to expose realistic system failure modes. Round 3: Grand Finale & Technical Evaluation (Online): Deliverables: Exported agent configuration files (architecture, prompt instructions, and system variables). Two complete audio call recordings showcasing successful task execution and failure-mode handling. Judging Criteria: Architecture & Prompt Engineering Data Schema & Variable Design Actions, Tool Calling & Intent Resolution Testing Rigour & Edge-Case Handling Call Quality, Latency & Conversational Naturalness