Ankit Neupane

I’m a software engineer specializing in backend systems and API infrastructure, with hands-on experience taking systems from design to deployment. I focus on reliability, scalability, and observable performance in production-ready services.

In my projects, I built a Semantic Cache Proxy that boosted LLM app performance by replacing string-matching with embedding-based retrieval. I designed a Redis-backed vector cache with TTL expiration, metrics for observability, and delivered strong results like a 70% cache hit rate and latency reduction.

I also built an end-to-end RAG pipeline prototype that ingests PDFs, generates embeddings, stores vectors, and retrieves semantically relevant chunks for grounded generation—without relying on frameworks like LangChain or LlamaIndex. Plus, I developed a production Notifier API with JWT authentication, RBAC, Redis rate limiting, and structured logging on a Dockerized Azure CI/CD setup.

Alongside backend work, I enjoy building user-facing systems too—recently engineering a media company website with top Lighthouse scores and a mobile-first React/TypeScript UI. I bring an engineer’s mindset to both AI and web systems: modular design, clean integrations, and outcomes you can measure.