Umar Areeb

I’m an AI Engineer specializing in LLM-based conversational systems, prompt engineering, and Retrieval-Augmented Generation (RAG). I focus on turning user needs into robust conversational flows, then validating real-world behavior through structured testing and iteration.

At Fiberisefit Pvt Ltd, I designed and iterated a conversational AI agent using prompt engineering and RAG across 2,700+ medical documents. By implementing self-querying retrieval, I improved response relevance by ~40% and strengthened safety guardrails and citation logic, achieving 100% compliance in medication and emergency refusal scenarios while improving response reliability.

I also evaluated and debugged system performance using 200+ structured test cases, improving factual accuracy, safety, and consistency. I defined nutrition-related user use cases and translated them into retrieval logic and conversational flows, aligning system behavior with real-world scenarios.

Previously, as a Prompt Engineer at Softage Information Technology Limited, I improved task completion by 25% through iterative prompt testing and refinement, using LangChain and RAG for contextual understanding in question answering and code generation workflows. Earlier, as an Associate Analyst at InfoEdge (99acres), I automated Python pipelines with Jenkins and Selenium, reducing manual effort by 40%, and I’ve also built analytical and machine-learning projects including trip analysis, SQL data mart insights, and crop disease detection with CNN. I hold a B.Tech in Electronics & Communication Engineering (2020–2024) and was a Winner in a College Ideathon tied to Accenture’s Data Analytics Virtual Experience (Forage).