AI Data Science Team is an open-source framework from Business Science that assembles a team of specialised AI agents to perform common data science work — loading and cleaning data, exploratory analysis, feature engineering and machine learning — with the goal of getting routine analysis done many times faster. Built on LangGraph and released under the MIT license, the project has collected more than 5,400 stars on GitHub and remained in active development in 2026.
Instead of one general-purpose chatbot, the framework splits data science work between agents with defined roles, coordinated as a LangGraph workflow: agents handle tasks such as data loading, data cleaning and wrangling, data visualisation, feature engineering and model building, passing results between each other and back to the analyst. The agents generate and execute Python code against your datasets, explain what they are doing, and can recommend next steps, which makes the system feel like a junior team of analysts working under your direction rather than a black box. Because it is a Python library, it plugs into notebooks and existing analysis workflows, and it works with OpenAI models as well as other providers. It is aimed at the repetitive 80 percent of data work — the loading, cleaning and first-pass exploration — so human data scientists can spend their time on interpretation and decisions.
Free. AI Data Science Team is open source under the MIT license; you only pay for the LLM API usage of whichever model provider you connect.
Data scientists, analysts and business teams who want AI agents to take on the repetitive loading, cleaning and exploratory stages of analysis inside their Python workflow.