LLM Engineer Intern — ArogoAI

Overview

During my internship at ArogoAI, I worked on developing an AI-powered Doctor Copilot system designed to assist with symptom triage and specialist diagnosis. My contributions focused on designing the end-to-end LLM architecture, enabling dynamic patient interaction workflows, integrating medical knowledge sources, and building retrieval systems to support clinically relevant diagnosis recommendations.

Key Contributions

• • Designed the end-to-end architecture of the AI-powered Doctor Copilot system, transforming patient symptom inputs into structured diagnostic insights using large language models.
• • Implemented embedding-based routing and integrated medical knowledge graphs to intelligently route cases to the most relevant medical specialist.
• • Built pipelines for dynamic follow-up questioning and symptom extraction, enabling interactive clinical reasoning and more precise patient assessment.
• • Generated longitudinal patient knowledge graphs from historical medical data to provide deeper insights into chronic conditions and recurring symptom patterns.
• • Performed LLM context engineering by optimizing prompt structures and memory strategies to maintain accurate and relevant outputs across multi-step patient conversations.
• • Integrated real-time clinical data sources including the ICD-11 API, enabling standardized medical classification within diagnosis workflows.
• • Built retrieval-based diagnosis pipelines using FAISS vector search with CrossEncoder re-ranking, along with structured logging pipelines to support reinforcement learning and system improvement.

Tech Stack & Tools

Letter of Recommendation

Letter of recommendation issued by ArogoAI leadership recognizing contributions to the development of the Doctor Copilot clinical AI system and LLM-based diagnostic workflows.