Zhang
Yecheng.
I build systems that help cities and AI reason better—from building-scale urban models to evaluation and alignment infrastructure.
清华大学博士生 · 字节跳动研究实习生
01 / Selected work
Research that moves between the city and the model.
Across urban sensing, multimodal alignment, and the infrastructure that makes AI systems more reliable.
VLM alignment · ECCV 2026
UrbanAlign
Training-free alignment between vision-language models and human preference.
Generative AI · Urban science
AI4UrbanScience
Evaluating and calibrating generative AI for scientific urban-data synthesis.
Open data · 32M+ buildings
Building-scale China
Open datasets and models that make the urban fabric measurable at national scale.
02 / Practice
BEIJING · 39.9042° N
116.4074° E
I work across two scales: how cities are understood, and how intelligent systems learn to understand.
Ph.D. Student · Tsinghua University
Building-scale urban analytics, simulation, spatial intelligence, and AI for urban science.
Research Intern · ByteDance
Evaluation, alignment, post-training, and self-improving infrastructure for large language models.