Divya Pratap Singh

I’m an AI Engineer focused on building and deploying production-grade Large Language Model (LLM) solutions, especially Retrieval-Augmented Generation (RAG) systems. I bring a strong backend foundation in Python and Go, designing scalable APIs and microservices with an emphasis on reliability in cloud-based enterprise environments.

At Celebal Technologies, I architect scalable RAG systems leveraging cloud-based LLMs and semantic frameworks, handling 5,000+ monthly queries with 90% response accuracy. I optimize retrieval pipelines with hybrid semantic ranking, improve context relevance by 25%, and reduce inference costs by 20% while cutting manual review by 40%. I also orchestrate asynchronous microservices for LLM endpoints to decrease latency by 20% (4.2s to 3.3s).

I enjoy turning complex requirements into measurable outcomes, from Kubernetes-backed Go services to enterprise GenAI assistants. I’ve delivered metadata-aware retrieval with vector search and document-level filtering (answer grounding +25%, citation precision +30%) and built an LLM-based code migration agent that migrated 120,000+ lines of legacy SAP ABAP into optimized PySpark pipelines with 87% syntax-level accuracy. My professional ethos is simple: improve system accuracy, latency, and cost efficiency—while keeping solutions maintainable with practical MLOps.