Recommendations & retrieval
Semantic search, similarity, and ranking — especially when language models enter the loop.
Machine learning · Vancouver
Xiang (Shawn) Liu
ML engineer working on recommendations, semantic search, and the space between experiments and production.
About
I build machine learning systems — mostly around recommendations, semantic search, and LLMs — and care about the gap between a good experiment and something that actually runs in production.
My background spans applied ML in industry and research in neuroimaging & medical ML at UBC. These days I think a lot about embeddings, evaluation, and making large models practical at scale.
MS Computer Science (ML) · Georgia Tech · MS Experimental Medicine · UBC · BSc CS & Statistics · UBC
On my mind 01 02 03 04
LLM systems
Data pipelines, inference efficiency, and evaluation for models in real products.
ML engineering
Bridging research and production: experimentation velocity, serving, and team workflows.
Applied research
Publishing and exploring ideas at the intersection of IR, ML, and real-world systems.
Papers
Contact
Happy to chat about ML systems, papers, or side projects.