Prepare the complete dataset and build the first working AI recommendation prototype by implementing:

✅ Phase 2 is considered complete only when the baseline model works end-to-end with mock data.

✅ Phase 2 Deliverables (Final Outputs)

Deliverable File Owner Status
Clean MovieLens Dataset ai/data/processed/ Elif x
Feature Tables (User/Genre) PostgreSQL Andaç ⬜
IBCF Baseline Model ai/models/ibcf.py Elif x
Offline Metrics Report ai/evaluation/offline_metrics.py Elif x
Mock AI API ai/serving/app.py Elif x
Mock Feed Endpoint backend/app/feed.py Berkay ⬜
iOS Feed UI (Mock Data) ios/Views/HomeFeed.swift Öykü + Can ⬜

ELIF — AI ENGINEER (Phase 2 Tasks)

Task Description Output File Status
Load MovieLens Read raw dataset ai/data/raw/ x
Clean Dataset Remove invalid ratings ai/data/processed/ x
Feature Engineering User–genre vectors feature_pipeline.py x
Implement IBCF Baseline recommender models/ibcf.py x
Offline Evaluation RMSE, Recall@K offline_metrics.py x
Serve Mock Recs Mock /recommend API serving/app.py x

🏗 BERKAY — BACKEND & DEVOPS (Phase 2 Tasks)

Task Description Output File Status
Create Mock Feed API Fake recommendations backend/app/feed.py ⬜
Connect to AI Service Proxy /recommend backend/app/feed.py ⬜
Auth Middleware Secure feed access backend/app/auth.py ⬜
API Response Format Match iOS needs docs/api_contracts.md ⬜
Local Docker Sync Backend ↔ AI infra/docker-compose.yml ⬜

🗄 ANDAÇ — DATABASE & DATA (Phase 2 Tasks)

Task Description Output File Status
Import MovieLens Into PostgreSQL database/seed.sql ⬜
Create Rating Tables Ratings & events Database ⬜
Create Feature Tables User-genre stats Database ⬜
Index Optimization Performance tuning Database ⬜
Verify Data Quality Nulls, duplicates Report ⬜