Shreyosi Endow

I earned my PhD in Computer Engineering at the University of Texas at Arlington, specializing in Human-Computer Interaction under the advisement of Dr. Cesar Torres. My research expertise spans natural language processing (NLP), crowdsourcing, generative AI (GenAI), and design research. My dissertation contributed automated and scalable techniques for extracting tacit knowledge from instructional content available on the web to better support tutorial search, generation, and navigation.

I currently work as an Applied Scientist contractor with the Scientific Evaluation team at Adobe, where I build large-scale evaluation sets for benchmarking generative AI models across imaging initiatives. As part of this role, I also develop automated pipelines that translate qualitative workflow insights gathered from creative professionals into quantifiable metrics that leadership can use to make model and feature adoption decisions.

Throughout my PhD, I was fortunate to participate in several research internships. At Adobe Research, I worked with Christine Dierk and Eunyee Koh to develop methods leveraging VLMs to extract and evaluate subjective design qualities like mood, enabling mood-based search and generation of design templates on Adobe Express. I also leveraged my background in wearables and sensing to contribute to Project Primrose, where I integrated motion-responsive behavior into a dynamic computational dress display. At Fujitsu Research, I worked at the intersection of HCI and computer vision, building a VLM-powered interactive system for evaluating the quality of large image datasets used for 3D reconstruction workflows.

I am always open to collaborations and happy to chat about research, creative tools, or new opportunities, so please do not hesitate to reach out!