DeepPrep: An LLM-Powered Agentic System for Autonomous Data Preparation

TL;DR

DeepPrep is an innovative LLM-powered system that autonomously constructs data preparation pipelines through iterative, execution-grounded interactions, enabling structured revisions and achieving high accuracy with lower inference costs.

Contribution

DeepPrep introduces a tree-based agentic reasoning framework and a progressive training method for autonomous, efficient, and accurate data preparation using LLMs.

Findings

Achieves data preparation accuracy comparable to GPT-5

Incurred 15x lower inference cost than GPT-5

Outperforms existing open-source baselines across datasets

Abstract

Data preparation, which aims to transform heterogeneous and noisy raw tables into analysis-ready data, remains a major bottleneck in data science. Recent approaches leverage large language models (LLMs) to automate data preparation from natural language specifications. However, existing LLM-powered methods either make decisions without grounding in intermediate execution results, or rely on linear interaction processes that offer limited support for revising earlier decisions. To address these limitations, we propose DeepPrep, an LLM-powered agentic system for autonomous data preparation. DeepPrep constructs data preparation pipelines through iterative, execution-grounded interaction with an environment that materializes intermediate table states and returns runtime feedback. To overcome the limitations of linear interaction, DeepPrep organizes pipeline construction with tree-based agentic reasoning, enabling structured exploration and non-local revision based on execution feedback. To enable effective learning of such behaviors, we propose a progressive agentic training framework, together with data synthesis that supplies diverse and complex ADP tasks. Extensive experiments show that DeepPrep achieves data preparation accuracy comparable to strong closed-source models (e.g., GPT-5) while incurring 15x lower inference cost, while establishing state-of-the-art performance among open-source baselines and generalizing effectively across diverse datasets.