Reproduce “Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Anoushka Vyas, Aarushi Dhanuka, Sina Khoshfetrat Pakazad, Henrik Ohlsson

DIA uses a coding agent in three roles—Data Interpreter, Schema Creator, and Query Generator—with shared workspace and memory. The Query Generator executes SQL and checks results against an output shape. Across seven benchmarks, it matches or exceeds prior results, including 55.7% task success on BIRD-Interact versus 22.7%. Evaluation covers only the Query Generator; execution checks cannot establish its interpretation is correct.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.