Back to Timeline
EL Nova Labs
Software Engineer
2024
Led a 5-engineer team building a React/Redux/GraphQL ERP serving 5,000+ concurrent users with 40% load time reduction. Shipped GPT-4-powered features and a LangChain RAG pipeline indexing 10,000+ documents at 85% retrieval accuracy.
## Role Overview
As a Software Engineer at EL Nova Labs, I led a 5-engineer team building a React/Redux/GraphQL ERP platform serving 5,000+ concurrent users, while also shipping AI-powered features that reduced administrative workload by 62%.
## Key Deliverables
- **React Platform at Scale** — Led a 5-engineer team building a React/Redux/GraphQL ERP for 5,000+ concurrent users. Drove 40% load time reduction via code splitting, memoization, and normalized state — zero production incidents post-launch.
- **AI & LLM Integration** — Shipped GPT-4-powered notice generation with voice input and a LangChain RAG pipeline indexing 10,000+ documents at 85% retrieval accuracy — cutting administrative workload by 62%.
- **Quality & CI/CD** — Established Vitest + RTL at 95% code coverage; automated blue-green GitHub Actions pipeline enabling zero-downtime releases across all production environments.
## Technical Approach
The ERP architecture used normalized Redux state with optimistic updates for real-time collaboration, code-split routes for sub-second initial loads, and a GraphQL gateway that batched queries to minimize waterfall requests. The RAG pipeline combined dense vector search with BM25 keyword retrieval for hybrid ranking, with automatic re-ingestion when documents changed.