Hi — I'm Zoey Zhang, a researcher focused on human-aligned, adaptive AI systems, NLP, knowledge graphs, and practical LLM applications. I hold a Master’s degree in Informatics (Data Science specialization) from the University of Zurich, and a Bachelor’s degree in Software Engineering. My academic and professional journey has always been guided by a simple drive: to connect advanced AI research with meaningful, real-world impact.
During my Master’s, I conducted research in Knowledge Graph Question Answering (KGQA), developing methods for text-to-SPARQL translation and answer verbalization for complex queries. Later as a Research Assistant, I built on this work to experiment with LLMs and interactive scholarly QA systems, producing prototypes and academic outputs. These experiences fostered my interest in building reliable, interpretable, and adaptive AI systems that can reason under uncertainty and bridge theory with practical applications.
In industry, I proposed and prototyped the GraphRAG project, which enhanced context retrieval and fact-consistency in deployed chatbot systems. While this experience focused on real-world deployment, it never shifted my academic interests, but strengthening my commitment to exploring reliable, interpretable, and adaptive AI systems. The project reinforced the importance of rapid prototyping, testbed-driven research, and iterative evaluation, all of which I now seek to combine with rigorous theoretical study in a PhD setting.
Alongside professional work, I have continuously pursued opportunities to advance my research expertise. For example, I attended ESSLLI Summer School 2025 and the RANLP Conference 2025 to deepen my understanding of reasoning, evaluation, and logic in NLP systems. These experiences further strengthened my commitment to safe, verifiable, and human-centered AI research, demonstrating that my academic interests have been consistently guiding my work—from foundational research to applied systems.
Research interest: I'm passionate about agentic AI that stays grounded and human-aligned. This means combining knowledge representation with RAG and LLMs to enable reliable, traceable reasoning—then adding human-in-the-loop refinement and explainable outputs so the system can safely support creativity, learning, and decision-making.
Skills: Python, Java, SQL · PyTorch · RAG pipelines · LLM prompt engineering · Neo4j & graph modelling · PostgreSQL · FastAPI / Django · Docker / Linux / Git · Autogen.
Outside of work and study I enjoy photography, reading and writing, and cooking. My interest in reading and writing naturally connects to my research: it motivates me to explore AI tools for knowledge organisation and AI-assisted writing—identifying unmet user needs and building prototypes that combine creativity with technical solutions.
For a complete overview of my projects and publications, please visit the Projects and Publications pages.
My primary creative outlet is programming, which allows me to transform ideas into tangible software and applications. Beyond research, I develop side projects such as a mobile MVP for shared memories and a web-based AI-assisted writing system. These projects help me explore human-centered AI, context-aware reasoning, and adaptive interaction, bridge theory with practice, and strengthen my problem-solving skills.🏃
Writing and reading are both creative outlets and tools for exploring knowledge. My interest in these areas motivates me to design AI tools for knowledge organization and writing assistance, combining technical skills with creativity.✍
I enjoy photography and cooking, which allow me to experiment, iterate, and refine ideas step by step. These hobbies reflect my curiosity, attention to detail, and enjoyment of creative processes. 📸