React ERP at 5,000 concurrent users
Who led the React ERP for 5,000 concurrent users at EL Nova Labs?
At EL Nova Labs, Rupam Das led five engineers on a React/Redux/GraphQL ERP for 5,000+ concurrent users, cut load time 40%, shipped LangChain RAG over 10,000+ docs at 85% retrieval accuracy, and held 95% coverage with zero-downtime releases.
Led a 5-engineer team on a React/Redux/GraphQL ERP; cut load time 40% via code splitting, memoization, and normalized state; shipped GPT-4 notice generation with voice input and a LangChain RAG pipeline over 10,000+ docs at 85% retrieval accuracy.
·
The constraint
An admin platform that is slow or untrustworthy at 5,000 concurrent users becomes unusable; document work that stays manual does not scale.
Architecture
Normalized Redux, route-level code splitting, and memoization for the 5,000-user ERP. Vitest plus React Testing Library at 95% coverage. Blue-green GitHub Actions for zero-downtime releases. GPT-4 notice generation with voice in the loop. LangChain RAG over 10,000+ documents with measured retrieval accuracy.
Sources / stack
- React
- Redux
- GraphQL
- GPT-4
- LangChain
- Vitest
- GitHub Actions
Scale and retrieval
- 01Client
- Code-split React
- Normalized Redux
- 02API
- GraphQL
- GPT-4 notices + voice
- 03Retrieval
- LangChain RAG
- 10,000+ docs
- 04Delivery
- 95% coverage
- Blue-green CI
What we measured
Load-time reduction
Code splitting, memoization, normalized state
Concurrent users
Zero production incidents post-launch
RAG retrieval accuracy
10,000+ docs indexed; 62% less admin workload
Test coverage
Blue-green GitHub Actions, zero-downtime releases
Result
40% faster loads; zero production incidents post-launch; 62% less administrative workload; 95% coverage; zero-downtime releases.