Welcome!

The world is increasingly driven by data, but turning large amounts of raw data into a few actionable insights often requires a team of well-trained engineers. The goal of the D2I lab (part of the Georgia Tech database group) is to 1) shorten this journey by improving both computational and human efficiency at every stage of the data lifecycle and 2) train world-class researchers along the way.

Our research focuses on building high-performance, user-friendly data systems to enable next-generation AI applications. Current research directions include:

  • Analytics over dirty and unstructured data
  • Data infrastructure for GenAI applications:

News

  • [Jan 2026] Our paper Stream2LLM has been accepted to MLSys'26.
  • [Dec 2025] Our paper VCR has been accepted to Information Systems'25.
  • [Nov 2025] Our paper HoneyBee has been accepted to SIGMOD'26.
  • [Jun 2025] Congrats to Peng Li for winning the 🏆 SIGMOD Jim Gray Doctoral Dissertation Award!
  • [Oct 2024] Our paper CanDE has been accepted to IEEE BigData'24.
  • [Sep 2024] Our paper Inshrinkerator has been accepted to SoCC'24.
  • [Sep 2024] Our paper Lotus won a 🏆 Best Paper Nomination in IISWC'24! Congrats Rajveer!
  • [July 2024] Our paper Lotus has been accepted to IISWC'24.
  • [July 2024] Dristi Shah received a VLDB 2024 Travel Award. Congrats Dristi!
  • [June 2024] Congrats to Peng Li for winning the 🏆 SIGMOD Research Highlight Awards at SIGMOD'24!
  • [June 2024] Kexin received a SIGMOD 2024 Distinguished PC Award.
  • [May 2024] SketchQL and VCR demo papers accepted at VLDB'24.
  • [Apr 2024] 🎓 Hantian Zhang defended his thesis! Congrats Hantian!
  • [Apr 2024] Kexin received an Amazon Research Award for optimizing dynamic layout designs in data analytics systems. Thanks Amazon!
  • [Apr 2024] Kexin received an NSF award to reimagine video moment retrieval with hand-drawn sketches. Thanks NSF!

Join

We are always looking for talented and motivated students who want to help push forward the agenda of democratizing data analytics. If you are a GT PhD student, please email us directly. If you are an undergraduate or master student, please fill out our research questionnaire.