Sepehr Asgarian is an AI Engineer at Chubb with more than six years of experience building applied machine learning and generative AI systems. He designs and delivers production-grade agentic assistants powered by large language models, retrieval-augmented generation, and MCP tooling, supported by the evaluation and observability frameworks needed to keep them reliable as models and requirements evolve. Before joining Chubb, Sepehr developed machine learning and recommendation systems across healthcare and fintech, translating applied research into deployed products. His expertise spans large language models, multimodal learning, and recommender systems, with a focus on bridging cutting-edge research and durable, production-ready engineering. A published researcher at NeurIPS, AAAI, and IEEE, Sepehr specializes in designing adaptable AI applications that reduce technical debt while delivering measurable business value in a rapidly evolving ecosystem.