Prompt Coding

Prompt Coding:

Ai-Assisted / Human-Centric Hackathon:

 Event Overview:

• This is a 24-hour, offline, high-intensity hackathon focused on problem-solving, system thinking, and engineering judgment.

The event is designed to simulate real-world constraints where participants must:

• Understand unfamiliar problems quickly
  
• Make trade-offs
  
• Build practical solutions
  
• Clearly explain and defend their work
  

This is not a long coding marathon.
It is a thinking-first, execution-second competition.    

2. Key Details:

• Mode: Offline
  
• Maximum Duration: 24hours
  
• Theme: Miscellaneous (open domain)
  
• Prize Pool: ₹60,000+
  
• Certificates: Participation certificates will be issued to all participants who complete the event.

3. Team & Participation Rules:

  Team Size:

• Solo participation is allowed
  
• Teams of up to 4 members are allowed
  
• Team size must not exceed 4
  

  College Restriction:

• Inter-college teams are NOT allowed
  
• All members of a team must belong to the same institute
  

4. Problem Statements:

Problem Release:

• Problem statements will be released on the spot
  
• No prior preparation, coding, or research is allowed
  

Source of Problems:

• If sponsors provide real or realistic problems, those will be used
  
• Otherwise, problems will be curated by faculty
  
• If no sponsor-specific problems exist, the domain remains miscellaneous
  

Problem statements will be:

• Clearly defined
  
• Open-ended
  
• Solvable within the event duration
  

5. Event Flow & Timeline (Indicative):

Exact timings will be announced on event day.

• Opening & Rules Brief
  
• Planning Phase:
  
  1. Understand the problem
     
  2. Define assumptions
     
  3. Design solution approach
     
• Core Build Phase – 
  
  1. Build MVP / working solution
     
• Feature Twist – Last Hour
  
  1. A new relevant feature or constraint is introduced
     
  2. Tests adaptability and design quality
     
• Final Demo & Explanation
  
  1. Live demo preferred
     
  2. Judges Q&A
     
• Results & Closing
  

Total time will not exceed 24hours.

6. Investigation Procedure:

•  Any judge or organiser may flag a team for investigation at any point during or after the event.
•  The flagged team is notified in writing and given 10 minutes to present their defence to the investigation
panel.
•  A panel of judges (none of whom evaluated that team) will deliberate privately and deliver a verdict within
30 minutes.
•  The event clock continues for the flagged team during investigation — no pause is granted.
•  Plagiarism rulings are final — they cannot be appealed under the standard appeals process

7. AI Usage Policy (Human-First):

1  Permitted AI Tools
•  LLM-based code assistants — ChatGPT, Claude, GitHub Copilot, Gemini, and equivalents — are fully
permitted.
•  AI image generation, design tools (Figma AI, Canva AI), and documentation generators are permitted.
•  AI may be used for debugging, code suggestions, test generation, documentation, and boilerplate
scaffolding.
•  AI-generated code is treated identically to human-written code — team members must be able to explain
it.
2  The Black-Box Rule
Over-reliance on AI that results in a system the team cannot explain, debug, or modify will be penalised. If a
judge asks a team member to alter a specific AI-generated function and they are unable to do so, that module
may receive a partial or zero score at the judge's discretion. The final system architecture, integration decisions,
and overall design must reflect the team's own understanding — not simply an AI output accepted without
review.
3  Strictly Prohibited Activities
•  Generating an entire end-to-end project from a single AI prompt and submitting it wholly unmodified.
•  Automated bots or scripts that commit code to the repository on behalf of team members without direct
human involvement.
•  DDoS attacks, ARP spoofing, packet sniffing, or any network interference targeting event infrastructure or
other teams.
•  Screen-sharing, remote-desktop access, or real-time code collaboration with individuals outside the
registered team. 
•  Using external services or APIs that are not publicly accessible (e.g., private paid APIs that constitute
pre-built logic for the PS)

8. Main Build Chat Rule (MANDATORY):

Each team must maintain one Main Build Chat, which serves as the primary AI interaction thread.

Main Build Chat:

• Used for:
  
• Architecture planning
  
• Core code generation
  
• Iterative improvements
  
• All AI-generated code included in the final solution must originate from this chat
  

Submission Requirement:

At the end of the event, teams must:

• Share the Main Build Chat link (or export, as instructed)
  
• This will be used post-event for verification if required
  

Side Chats:

• Allowed for:
  
• Brainstorming
  
• Learning concepts
  
• Clarifying doubts
  
• AI-generated code from side chats must not be directly copied into final solution
  

Manual (human-written) code is allowed.

9. Prompt Budget System (important – Read Carefully):

Purpose:

The prompt budget exists to:

• Encourage thinking before prompting
  
• Prevent AI spamming
  
• Add strategic depth
  
• Keep competition fair
  

How It Works:

• Each team is given a fixed prompt budget
  
• Prompt budget will be announced on the spot
  
• The number disclosed during the event is final
  

 Example only (NOT final):
If a team is given 10 prompts, it means:

• They may send 10 meaningful messages to the Main Build Chat
  
• Each meaningful request for code, logic, or design counts as oe prompt
  

This number is only an example.

Feature Twist Prompt Budget:

• When the final feature/constraint is introduced:
  
• Teams will receive additional prompts
  
• This ensures fairness and allows adaptation
  

Example only (NOT final):

• If initial budget = 10 prompts
  
• Feature twist bonus = +3 prompts
  
• Total = 13 prompts

Again, numbers are examples.

What Counts as a Prompt:

• Any message intended to:
  
• Generate code
  
• Modify logic
  
• Design architecture
  
• Minor clarifications may be ignored at organiser discretion
  

What Happens If You Exceed:

• Teams exceeding the budget may:
  
• Lose points
  
• Be penalized during judging
  

Details will be announced clearly before coding starts

10. Judging Criteria (5-Point Scale):

Each criterion is scored 1 to 5.

1. Problem Understanding & Framing
   
2. Novelty & Approach
   
3. Engineering Decisions & Design
   
4. Execution & Demo Quality
   
5. Explanation & Clarity:
   

Judges will not judge:

• Lines of code
  
• AI vs manual ratio
  
• Tool complexity
  

11. Explanation-Based Evaluation (Core Fairness Rule):

Every team must:

• Explain their system architecture
  
• Justify design choices
  
• Answer follow-up questions
  
• Respond to a “what if we change X?” scenario
  

Failure to explain clearly may result in:

• Score reduction
  
• Disqualification in extreme cases
  

12. Fully AI-Generated Solutions:

• Solutions built entirely with AI assistance are allowed
  
• Such teams must:
  
• Demonstrate full understanding
  
• Defend decisions
  
• Adapt to questions
  

Understanding determines score — not code origin.

13. Submissions:

Each team must submit:

• Project code (GitHub or ZIP)
  
• Main Build Chat link/export
  
• Short explanation/demo (live)
  

Exact submission format will be shared during the event.

14. Code of Conduct:

• Fair play is mandatory
  
• Misrepresentation or dishonesty may lead to penalties
  
• Respect participants, judges, and volunteers
  
• Organisers’ decisions are final
  

15. Results, Prizes & Certificates:

• Winners will be announced at the end of the event
  
• Prize distribution will be disclosed before judging begins
  
• All participants receive participation certificates
  

16. Final Notes:

• All remaining logistical details will be disclosed before the event
  
• Organisers reserve the right to:
  
• Modify timelines slightly
  
• Clarify rules
  
• Resolve edge cases
  

The spirit of the event is fair competition, learning, and clear thinking.

Price pool will be subject to the number of participants if number of participants is not adequate the competition will get cancelled and the money will be refunded to you bank account