Research role focused on building an evaluation framework for AI coding agents to improve developer experience on the Modal platform. This involves defining quantitative objectives, measuring performance, and translating insights into product improvements, while staying updated on agent advancements and customer use cases.