About the lab
DAISY — Data-centric AI Systems — is a research lab in the Department of Computer Engineering at Sharif University of Technology.
DAISY Lab works where machine learning meets data systems. We ask how learning can make databases and data infrastructure faster and easier to use, what infrastructure machine learning itself needs to work reliably in production, and how to reason about the networks that data lives in.
Most of our problems come from real deployments. The lab’s director has worked in industry on machine learning infrastructure, business intelligence and large-scale systems, and the lab keeps that habit: we choose problems because they matter outside the paper, and we expect results to survive contact with real data.
How we work
- Data first. In our problems the bottleneck is usually the data — its schema, its quality, what is missing from it — not the size of the model.
- Systems that can say “I don’t know”. Calibration, uncertainty and verification run through all of our work.
- Open by default. Code and data are released with papers whenever we can.
The name
A daisy is one flower made of many small ones. The lab’s research directions are its petals, and data sits at the centre.