Official Documentation

Hackathon Playbook

Agentic AI Challenge

This hackathon evaluates real agentic AI system design, reasoning, and implementation. Move beyond no-code automations toward thoughtful, explainable agent architectures.

Development Rules

⚠️ Very Important - Read Carefully

❌ NOT Allowed

No-code or low-code automation tools are strictly prohibited:

n8n
Zapier
Make (Integromat)
Any drag-and-drop workflow builders
🚫

These tools abstract away core reasoning and decision logic, which defeats the purpose of this challenge.

Recommended Tech Stack

Code-first approaches preferred

Build using code-first approaches, preferably in Python, with clear and explainable agent architecture.

Agent Frameworks

LangChain
LangGraph
CrewAI

LLMs

LLaMA
Mistral
Phi
Gemini
Open Source

APIs & Tools

  • Public/free APIs
  • File system tools
  • Email/messaging APIs
  • Free tiers only

Note: Other open-source frameworks allowed if agent logic is clearly implemented and explained.

AI Coding Assistants

Allowed with conditions

Allowed Tools

Cursor
Copilot
Antigravity

⚠️ Critical Rules

  • Use for assistance only
  • Must explain all logic
  • No blind copy-paste

This is NOT a "prompt and ship" challenge

Unexplainable AI-generated code may result in disqualification

GitHub & Code Monitoring

Continuous evaluation

Before Start

Share an empty GitHub repository with organizers

During Event

  • Regular monitoring
  • Frequent commits

Evaluation

  • • Commit history
  • • Code quality
  • • Evolution

Good Practices

  • Incremental development
  • Clear architectural decisions
  • Refactoring & improvements
  • Thoughtful commit messages

Bad Practices

  • One large code dump
  • Vibe coding without structure
  • Unexplained generated code

Problem Statement Allocation

First come, first serve

5️⃣

5 problem statements available

🎯

Limited slots per statement

First come, first serve allocation

🔄

Choose another if full

This ensures diversity of solutions and fair evaluation

Evaluation Deliverables

Three-stage comprehensive review

The hackathon features three stages of evaluation to ensure a comprehensive review of all projects.

1

First Evaluation

3:00 PM - 6:00 PM

Focus: Project conceptualization and initial setup

DeliverableRequirement
Problem understandingA brief description of the understanding of the problem being solved
Proposed Solution/ApproachInitial thoughts and relevance of Solution
🎵

Break Time

6:30 PM - 9:00 PM

Teams can attend the concert and go through the stalls for snacks near swimming pool area.

2

Second Evaluation

1:00 AM - 4:00 AM

Focus: Technical execution and progress towards working prototype

DeliverableRequirement
Tech stack/architectureDescription about tools used
Implementation of proposed solutionPrecision of Solution with respect to proposed solution
3

Final Evaluation

10:00 AM - 12:30 PM

Focus: Comprehensive submission and live presentation for top teams

Submission Requirements (9:00 AM - 10:00 AM)

DeliverableRequirementFormat
Working Product (Bonus)Deployed linkURL
Demo VideoA max 2-minute video demonstrating the project's featuresVideo file
Presentation DeckSlides outlining the problem, solution, technology, and future scopePDF/PPT

Top 10 Team Presentation

12:30 PM - 3:00 PM

The top 10 teams will be required to present their projects live to the panel of judges.

Demo

5 minutes for a live demonstration of the solution

Q&A

5 minutes of questioning by the judges

🏆

Winner Announcement

3:00 PM - 3:30 PM

Main Stage

The judges' decision is final.