At York University, I investigated machine learning techniques for classifying professional emails and co-developed routing tables for autonomous email pipelines. The work explored sending simple queries to local retrieval systems and complex prompts to large language models.
I built a multi-source agentic RAG system that routes queries across document retrieval, web search, Wikipedia, and external APIs. Its hybrid retrieval pipeline achieved 1.000 RAGAS Context Precision while maintaining 1.000 Context Recall, and I deployed the FastAPI service on AWS ECS Fargate.
I’ve also developed multi-agent projects for travel planning and disruption recovery, research assistance, and conversational RAG. My hands-on work spans LangGraph and Google ADK, Azure data pipelines, and cloud-native applications.

