Using LLMs for data engineering in enterprises is challenging
Large language models (LLMs) promise to automate data engineering on tabular data, offering enterprises a valuable opportunity to reduce costly manual work. But adopting LLMs in industry brings unique challenges, such as large table sizes and the need for internal knowledge. Our analysis shows that LLMs face substantial limitations in real-world scenarios, and offers ideas to support better adoption of LLMs in industry.
Authors
- Jan-Micha Bodensohn (ORCID: https://orcid.org/0000-0003-4884-0300)
- Liane Vogel (ORCID: https://orcid.org/0000-0001-9768-8873)
- Anupam Sanghi (ORCID: https://orcid.org/0000-0003-4764-3583)
- Carsten Binnig (ORCID: https://orcid.org/0000-0002-2744-7836)
- Ulf Brackmann
Publication Details
- Journal
- Publication Server of Goethe University Frankfurt am Main (Goethe University Frankfurt)
- Published
- 2026-01-05
- Primary Topic
- Machine Learning and Data Classification
- Type
- article
- Field-Weighted Citation Impact
- 0.00