About

I’m YangCC, the public identity I use for technical work and conversations with the developer community.

My current focus is multimodal time-series forecasting: understanding how signals from different modalities can be aligned, represented, and used to model temporal dynamics. I’m especially interested in forecasting systems that remain useful when observations are noisy, incomplete, or distributed unevenly over time.

Alongside that research direction, I explore vector database systems and the retrieval infrastructure behind modern AI applications—from approximate nearest-neighbor search and indexing to the behavior of end-to-end retrieval pipelines.

Current interests

• Multimodal representation learning for temporal data
• Forecasting under missing, noisy, and heterogeneous signals
• Vector indexing, approximate nearest-neighbor search, and retrieval systems
• Reproducible experiments and clear technical communication

This site

This is where I publish working notes and technical articles. Selected project case studies will appear when they are ready. The writing may be in English or Chinese, depending on the subject and the community it is intended for.

Contact

Technical discussions, open-source collaboration, and professional opportunities are welcome. You can find me on GitHub or reach me by email. A dedicated contact page is also available at Contact.