Ghulam Jillani

I've delivered production Generative AI, LLM, and RAG systems for Nuclieos, Crypto Express, Techlogix, and remote contract clients, spanning AWS, GCP, and Azure. My work turns prototypes into reliable AI products with measurable improvements in cost, latency, uptime, and release speed.

In recent contract work, I shipped a LangGraph, FastAPI, and Kubernetes multi-agent service that cut manual operations effort by about 55%, and built a multi-provider LLM gateway that reduced LLM spend by about 40%. I also migrated a workload to LoRA fine-tuned Llama 3 on vLLM, lowering inference costs by about 35% at equal task quality.

At Nuclieos, I built LLM and RAG pipelines with LangChain, LlamaIndex, and LangGraph, reduced query latency by about 40%, and established multi-cloud MLOps infrastructure sustaining 99.9% uptime. I led six engineers and data scientists, introduced governance and observability practices, and shortened release cycles from weeks to under 48 hours.

I enjoy solving enterprise AI delivery challenges through practical architecture, evaluation, guardrails, vector-backed retrieval, and scalable deployment. I'm open to fully remote contract, freelance, advisory, and consulting engagements.