At MRCC Group, I decomposed a monolithic Enterprise LMS into modular microservices, increasing deployment velocity by 40%. I also built a LangChain and Redis vector-store RAG pipeline that reduced Tier-1 support queries by 30%.
I cut LLM token spend and p95 latency by 70% with Redis semantic caching, and helped sustain 99.9% production reliability through testing, CI/CD, logging, and observability improvements.
As an independent software engineer, I’m building a LinkedIn content and scheduling platform with FastAPI services, a fault-tolerant publishing pipeline, and an LLM content engine that uses semantic retrieval over users’ post histories.
