PayLink: Building a Multi-Database Backend

PayLink is a simplified peer-to-peer payments backend where Supabase, MongoDB, and Neo4j each handle exactly what they're best at — authentication, flexible transaction receipts, and relationship/graph tracing — all sitting behind a single FastAPI service. I walk through the full architecture, show it running live, and I'm upfront about the trade-offs too — because no architecture is free, and understanding what you're giving up matters just as much as what you're gaining. 🔗 Full source code, architecture diagram, and README: https://github.com/saurabhkamal/Desig... In this video: ✅ Why a fintech app genuinely needs three databases, not one ✅ Supabase for JWT authentication ✅ MongoDB for flexible, schema-less transaction receipts ✅ Neo4j for multi-hop payment relationship tracing (with a live Cypher demo) ✅ How one endpoint safely writes to two databases in a single request ✅ The security gap this closes — and why it matters ✅ Honest trade-offs of a multi-database architecture 🛠️ Tech stack: FastAPI, Supabase (Auth + Postgres), MongoDB, Neo4j (Community Edition), Docker, Python, Pydantic If you're prepping for backend engineering interviews, I think the real value here isn't "I used three databases" — it's being able to explain why, and being honest about what it costs. That's what I tried to show in this video. If this was useful, consider subscribing — more backend engineering, database design, and system architecture content is on the way. #FastAPI #MongoDB #Neo4j #Supabase #BackendDevelopment #SystemDesign #Python #GraphDatabase #Fintech #SoftwareEngineering #Docker #API