Data-centric foundations
We create the data, representations, and evaluation foundations that make AI systems more capable, reliable, and ready for consequential problems.
City University of Hong Kong · DSAIL Lab
We build AI with frontier techniques, grounded in high-quality data, to expand scientific breakthroughs and beyond.
Based in the Department of Data Science and the Hong Kong Institute of AI for Science.
DSAIL
Lab news
DSAIL officially launches at City University of Hong Kong.
We are recruiting founding members at every level — PhD students, postdocs, RAs, and student collaborators. See open positions
"Graph Diffusion Transformers are In-Context Molecular Designers" was accepted at ICLR. Paper
torch-molecule, our open-source toolkit for molecular discovery, was released. GitHub
Research
We pursue foundational questions, validate them on consequential scientific problems, and turn strong ideas into open systems.
We create the data, representations, and evaluation foundations that make AI systems more capable, reliable, and ready for consequential problems.
We advance foundational AI approches that can reason, learn, and work with the complexity of the real world.
We use AI to make an impact and accelerate discovery, while translating scientific intelligence into systems that expand what people can create and understand.
01 · Data-centric foundations
02 · Core AI research
03 · Science and beyond
People
Early members receive close mentorship, meaningful ownership, and the freedom to shape the lab's research agenda.
Open position
PhD researcher
Own a bold research direction from first principles to publication and open-source impact.
Explore the opportunityOpen position
RA or student collaborator
Turn technical depth and research curiosity into work that matters.
Explore the opportunityJoin DSAIL
DSAIL is for self-directed researchers who want to develop deep technical expertise, own ambitious ideas, and produce work with lasting scientific impact. We welcome talent at every level — from the most experienced (postdocs) to beginners with strong interests (student collaborators).
Postdocs
Lead ambitious research with real independence and full support for career development.
Details & applyPhD students
Fully funded four-year positions every year, with freedom to explore, flexible mentoring, full support for career development, and diverse admission pathways.
Details & applyResearch assistants
Build the data, training, evaluation, and research infrastructure that makes frontier AI work in practice.
Details & applyStudent collaborators
Work closely with us on a focused project and grow toward independent, publication-quality research.
Details & applyContact
We welcome conversations with researchers, students, and collaborators from both academia and industry working at the frontier of AI and scientific discovery.