Phase 1: Exploration & Ideation
Resource discovery, team formation, and strategic planning
1
Kickoff & Resource Exploration
- Explored free resources and perks from hackathon
- Free domain (Entri), API credits (Algorand/IPFS, Dappier)
- Platform credits (RevenueCat, Bolt.new, Pica, Sentry, Expo, Lingo, Tavus)
- Realized Bolt.new requirement for final product
2
Team Formation & Ideation
- Richard Constantine joined as Product Manager
- Explored multiple concepts: meeting consultant platform, OCR invoice processor
- Investigated due diligence automation and CTO services platform
3
Strategy Inspiration
- Reviewed marketing strategy via FigJam
- Studied River co-founder insights
- Analyzed one-week product case study and "35 Playbook"
4
Team Alignment
- First full-team sync with complete roster
- Muhammad Qasim (Frontend), Richard Constantine (PM)
- Timothy Adetunji (Frontend), Piotr Żak (Backend)
- Analyzed 36 hackathon judges for strategic alignment
- HR tech and cold outreach identified as go-to-market strategy
5
Prioritization & Market Fit
- Validated ideas with marketing/sales expert R N
- Explored monetization potential vs. social impact
- Shortlisted: AI-driven cost optimization, regulatory tools, collaborative storytelling
- Emphasized feasibility, scalability, and post-hackathon monetization
6
Solution Chosen 🎯
- Created comprehensive idea-exploration page
- Team vote selected: Marketing automation platform for social influencers & YouTubers
- Meeting recap documented on Otter
- Market analysis: 51M influencers, 26% YoY growth
- Addressed feasibility and market saturation concerns
Phase 2: Development & Foundation
Engineering begins, architecture decisions, and core development
7
Engineering Begins
- Developed first code: Multi-Agent System API for YouTube agents
- Automated video aggregation by author
- Initial deployments and team role alignment
8
Sales Funnel + PRD
- Discussed sales funnel and automation via n8n
- Richard prepared Product Requirement Document (PRD)
- Identified early competitors (e.g., Aspire.io)
- Defined AI stack: NLP, CLIP, GPT, sentiment detection, forecasting
9
Backend Foundation
- Created GitHub repo and onboarded developers
- Developed API endpoints using Clean Architecture
- Agreed to use Kaggle datasets for MVP
- Finalized tech stack: Firebase, Next.js, Python, BigQuery
10
Feature Breakdown
- Created GitHub board with three modules:
- • AI Matching Engine
- • Campaign Setup & Brief Generation
- • Content Ideation Assistant
- Continued implementation and dataset integration
11
Frontend Setup
- Built early frontend with Lovable, v0.dev, Bolt
- Aligned data scientist processes with BigQuery
- Setup cloud deployment
- Defined backend-frontend contracts
12
Infrastructure Issues & Team Changes ⚠️
- Time zone differences created communication delays
- Significant backend progress (brand models, matching)
- BigQuery tables disappeared - global outage (DownDetector)
- Timothy left team, recruitment initiated
- New marketing/sales expert from U.S. joined
13
Data Modeling
- Created seed data for fictional brand ecosystem:
- • Authentic Scope, Quality Care, Synthapse
- Used nested JSON in BigQuery for scalable modeling
- Finalized contract interface for frontend integration
14
Frontend Progress
- Frontend deployed: https://ai-creater-app.vercel.app/
- Functional components: Brand onboarding, Campaign editing, AI Clones
- Conducted alignment meeting and ticketed key tasks
- Strengthened team cohesion despite recent changes
15
MVP Functionality 🚀
- Sunday light work session (~2h)
- MVP of AI Matching Engine completed (vector embeddings-based)
- User email handling added to brand data model
- Pull requests submitted for review
- Improved workflow and team process maturity
Phase 3: MVP & Integration
Advanced features, video analysis, and system integration
16
Planning & Ticket Assignment
- No code committed - strategic planning day
- Assigned Ticket #11: Generate AI-driven campaign brief via Raport.API
- Set objective to generate campaign summaries based on brief inputs
17
Core Modules Advancing
- Refactored and optimized AI Matching Engine
- Implemented backend for AI-powered brief generation
- Started AI Insights / Trend Detection implementation
- Architecture remains clean and modular under Synthapse ecosystem
18
Video Analysis Exploration
- Focused on Video Analysis pipeline using TikTok videos
- Encountered OpenCV installation challenges (MacOS arm64 vs x86_64)
- Resolved using Rosetta
- Began testing frame extraction from videos via URL input
19
New Deployment & MVP Scope Set
- Deployed new system version with working modules
- Processed first batch of 100 TikTok influencers
- Team aligned that 1,000 influencers suffice for MVP scope
- Added "Refinement" column to ticket board for clarification
20
Scoring & Virtual Influencers
- Added scoring logic to Matching Engine (match %)
- Started Virtual Influencer Module for brand campaigns
- Merged large frontend/backend branches after review
- Scheduled refinement meeting to address mismatches
21
Refinement & Planning
- Held full-team refinement meeting
- Walked through UI components
- Identified backend/UX areas needing attention
- Clarified engineering priorities and simplified brand campaign logic
22
Rest Day
- No hackathon work today (Sunday)
23
LinkedIn Integration
- Integrated LinkedIn OAuth login - working locally with callback
- Discussed redirecting domain DNS to Cloudflare
- Decided to store LinkedIn user data in Firebase
- Considering frontend work for LinkedIn onboarding next
24
Frontend Progress & Backend Alignment
- Significant progress on frontend-backend integration
- Successfully receiving brand onboarding data
- Displaying influencer information
- Addressed deployment pipeline issues
25
Frontend User Profile
- Running frontend project locally and added Redux
- Managing global React state for user brand
- LinkedIn Integration research
- Understanding CSRF attack prevention
Phase 4: Refinement & Deployment
Final integration, deployment, and project completion
26
Frontend LinkedIn Integration
- Refinement meeting with team
- Decided Sunday is final deployment + code freeze
- LinkedIn Integration complexity discovered
- DNS repointing from Cloudflare to Vercel
27
Brief Generation
- Frontend work around campaign and brief generator
- Need to rewrite from .jsx to .tsx for better maintainability
- Writing PR for another programmer
- LinkedIn Integration issues due to developer platform access
28
Organize Code/Deployment
- Organizing repository and strengthening processes
- Resolving deployment errors
29
Integrating Developer Work
- Rest day - minimal hackathon work
- Merging dev into main branch
30
Future Possibility of Expanding 🎉
- Team gathering and detailed discussions
- Recording demos for submission
- Preparing sales pitch
- Project submission completed
- Discussing future opportunities