What's included

• Full-Stack AI Web App (Next.js)

  Ship a production Next.js app with streaming chat UX, dashboards, and integrated AI features (ready for real users, not a demo).

• Agentic Workflow

  Deliver a supervisor+specialists agent that can call tools (DB/search/graph) with guardrails, retries, and structured state.

• Production RAG Assistant (End-to-End)

  Build a grounded LLM assistant over your docs (ingestion → chunking → embeddings → retrieval → generation) with citations and safe fallbacks.

• Multi-Strategy RAG + Quality Benchmarking

  Implement and compare multiple RAG approaches (hybrid/graph/corrective) and ship an eval harness (RAGAS-style metrics, regression tracking).

• GraphRAG (Neo4j + Vector Search)

  Add entity extraction + knowledge graph traversal alongside vector retrieval for multi-hop questions and better relationship reasoning.

• Secure Auth + Multi-Tenant Data Access (Supabase/RLS)

  Design the database model and row-level security for safe multi-tenant apps, including roles/permissions and protected APIs.

• FastAPI Backend for AI + Async Jobs

  Build a clean FastAPI service for AI endpoints, background tasks, and integrations (high-concurrency patterns where needed).

• Offline-First PWA Module (Sync Engine)

  Implement offline capture + queued writes + conflict-safe sync (service worker + IndexedDB) for field/clinical-style workflows.

• Cloud Deployment on GCP + CI/CD Hardening

  Containerize and deploy to Google Cloud with secrets/config management, environments, and a release pipeline with quality gates.