1. Project Overview Optima AI is a hybrid, privacy-first supply chain intelligence platform. This is NOT: a simple dashboard a CRUD app a basic SaaS tool This is a system that includes: local execution (Edge Agent) cloud platform (multi-tenant) data pipelines planning & optimization engine AI-assisted decision layer 2. System Architecture (High-Level) The system consists of: 1. Edge Agent (Client-side) Runs in customer environment Handles: data ingestion validation transformation planning calculations alerts generation recommendations 2. Cloud Platform Multi-tenant SaaS Handles: authentication dashboards configuration sync APIs AI orchestration 3. Sync Layer Secure communication between Edge and Cloud Payload-based (no raw data required) 4. AI Layer Works on structured outputs Generates summaries and explanations 3. Tech Stack (Mandatory) Backend Python (FastAPI) PostgreSQL Redis (for jobs/caching) Celery or RQ (async processing) Frontend Next.js TypeScript Tailwind CSS Charting (Recharts / ECharts) Deployment Docker (required) REST APIs only (no GraphQL) 4. Core Technical Challenges (What we are solving) Freelancers must be comfortable with: 1. Hybrid architecture part of system runs locally part runs in cloud sync layer must be reliable and secure 2. Data ingestion pipeline messy Excel / CSV inputs validation and normalization schema mapping (config-driven) 3. Planning & analytics engine inventory projections demand vs supply logic KPI calculations rule-based alerts 4. Config-driven system no hardcoded logic per client mapping + thresholds configurable 5. Multi-tenant system strict organization isolation scalable backend structure 6. Async job processing ingestion jobs calculation jobs sync jobs 5. Module Breakdown Edge Agent ingestion module validation module mapping module planning engine alerts engine recommendation engine sync module Cloud Backend auth module organization module sync API dashboard API alerts management settings/config AI orchestration Frontend dashboard UI alerts center settings AI chat recommendations panel 6. Data Flow (Simplified) Client uploads data locally Edge Agent: validates transforms computes KPIs & alerts Edge sends structured payloads to Cloud Cloud: stores summaries renders dashboards AI: reads structured outputs generates insights 7. Optima Recommendations+ (Key Feature) This is a structured recommendation system: Requirements: generate recommendation objects from calculations allow adding advisory notes support weekly insights generation Output must include: risk root cause recommended action priority business context alignment 8. Developer Expectations You should: understand system design (not just coding) write modular, clean code handle data-heavy logic design APIs properly think about scalability 9. What We DO NOT want Do NOT apply if you: only build dashboards rely on no-code tools prefer quick hacks hardcode logic per client don’t understand async systems 10. Deliverables production-ready code modular architecture Dockerized services API documentation clean database schema basic logging and error handling 11. Evaluation Criteria Candidates will be evaluated on: architecture understanding ability to explain hybrid systems data pipeline experience API design quality code structure