docs/roadmap.md

HearthNet Hackathon Master Plan

Timeline: June 6 → June 15 Theme: Community-owned resilient AI infrastructure

Team Structure Role Person Responsibility Product Lead Christof Vision, architecture, scope Tech Lead TBD Technical decisions Backend Engineer TBD Capability Bus AI Engineer TBD LLM & Routing RAG Engineer TBD Knowledge system Frontend Engineer TBD UI DevOps Engineer TBD Deployment Marketing Lead TBD Social, branding Pitch Lead TBD Presentation Community Manager TBD Discord, recruiting Overall Roadmap Date Product Engineering AI DevOps Marketing Pitch Jun 6 Team Formation Repo Setup Research Infra Setup Recruitment Story Outline Jun 7 Architecture Freeze Core Framework Model Selection Docker Branding Problem Statement Jun 8 MVP Scope Discovery LLM Node CI/CD First Post Demo Flow Jun 9 Review Routing RAG Containers Screenshots Narrative Jun 10 Scope Check Integration Retrieval Deployments Progress Post Draft Slides Jun 11 Testing Features Optimization Monitoring Community Feedback Pitch Draft Jun 12 Feature Freeze Bug Fixes Bug Fixes Bug Fixes Video Clips Presentation Jun 13 Demo Review Polish Polish Polish Social Push Demo Script Jun 14 Final Review Final Fixes Final Fixes Final Fixes Launch Post Rehearsal Jun 15 Submission Submission Submission Submission Promotion Presentation Product & Architecture Track June 6 Deliverables Confirm HearthNet concept Define MVP Define stretch goals Create architecture diagram Create user journey Questions to Answer What problem are we solving? Why now? Why is local AI important? What does success look like on June 15? June 7 Architecture Freeze Discovery mechanism Capability Bus Routing layer RAG integration Frontend architecture

After this date:

❌ No major redesigns

Engineering Track Backend Week 1 Task Priority Node Registry Critical Discovery Critical Capability Announcements Critical Routing Engine Critical Health Monitoring High Service Registry High Metrics Medium Checklist Nodes discover each other Nodes advertise capabilities Requests routed correctly Failover works AI Track Local Inference Research llama.cpp Ollama vLLM MLX (optional) Tasks Select model Benchmark model Create inference endpoint Remote execution Stretch Multi-node inference Federated learning concept RAG Track Build Emergency Knowledge Library

Potential datasets:

First aid Emergency response Community handbook Local resources Tasks PDF ingestion Chunking Embeddings Retrieval Citation support

Demo question:

"How do I perform CPR?"

Should work even with internet disconnected.

Frontend Track Screens Screen Priority Network Map Critical AI Chat Critical RAG Search Critical Node Status High Marketplace Medium Community Chat Medium Visual Goal

Judges should understand everything in 10 seconds.

DevOps Track Infrastructure Tasks Docker Docker Compose GitHub Actions Logging Monitoring Demo Setup

Minimum:

Laptop A Laptop B Raspberry Pi

Preferred:

3 laptops 1 Raspberry Pi Marketing Track

This is where most hackathon teams completely fail.

June 6 Recruitment Post Discord post HF discussion Reddit post Team call invitation

Goal:

Find engineers.

June 7 Branding Name finalized Logo Banner Color scheme Elevator pitch

Example:

HearthNet: Community-Owned AI That Works Even When The Internet Doesn't.

June 8 Landing Page Simple website Problem statement Architecture image Team section June 9 Content

Create:

Architecture diagram Screenshots GIFs Progress updates June 10 Community Feedback

Ask:

Would you use this? Why? What is missing?

Gather comments.

June 11-13 Daily Posts

Post:

Progress Screenshots Team photos Failures Lessons learned

Judges often recognize active teams.

June 14 Launch Post

Show:

Demo Screenshots Architecture

Create buzz before judging.

Pitch Track Day 1

Create story.

Problem

Current AI:

Centralized Fragile Requires internet Solution

HearthNet

Local-first Community-owned Resilient Day 2

Demo Narrative

Scene 1

Internet working

Ask AI question

Works

Scene 2

Pull network cable

Audience sees disconnect

Scene 3

Ask same question

Still works

Audience understands value instantly.

Day 3

Future Vision

Phase 1

Neighborhood

Phase 2

City

Phase 3

Federated Communities

Phase 4

Global Community AI Network

Daily Standup Template

Every day:

Question Answer What did I finish? What am I doing today? What's blocking me? Do I need help?

Maximum 15 minutes.

Critical Milestones Date Milestone Jun 6 Team Formed Jun 7 Architecture Frozen Jun 8 Discovery Working Jun 9 LLM Running Jun 10 RAG Working Jun 11 End-to-End Demo Jun 12 Failover Working Jun 13 Feature Freeze Jun 14 Pitch Rehearsal Jun 15 Submission