Skip to content

Hire your first AI data scientist
as a new member of the Team

Developed by the scikit-learn team, Skore is the statistical harness for data science agents that builds, validates, and tracks models and applies the methodology of the scikit-learn maintainers at every step. It is deployed on-prem or inside your own cloud, at a fixed price and scope. Your team builds the models, Skore makes sure they hold up.

Book a call

  • Meet Skore

    An AI teammate with methodology built in

    Skore is the statistical harness that runs inside the coding tools your team already uses, like Claude Code, OpenCode, or GitHub Copilot. It doesn’t replace your data scientists. It handles the repetitive legwork: exploring data, building pipelines, evaluating results with the statistical rigor of scikit-learn’s methodology baked into every step.

    • Every experiment lands in the Evaluation Hub: ranked, comparable, and traceable. Your team reviews and decides. The agent does the legwork.

    Skore collage: a data scientist on the phone by a window, next to custom check code registered as a reusable skill.

  • Private by design

    Your stack. Your cloud.
    Your data stays yours.

    Skore deploys inside your own VPC on AWS or Google Cloud: No data leaves your boundary. It works with the frameworks your team already uses: scikit-learn, XGBoost, TabICL, and it connects to your data sources (BigQuery, DataBricks, Snowflake or CSV/Parquet files).

    • There’s no migration, no lock-in, and no new infrastructure to maintain. Skore rides alongside what’s already in place.

    Skore collage: two colleagues talking at a desk in an office, next to a summary of findings from 16 automated checks.

One license.
Everything included.

Skore Team comes at a fixed price of €25,000 or $30,000 per year.
The license covers:

  • A woman standing by a colleague’s desk in an office, talking with him as he looks up from his laptop.

    Installation

    Deployed on-prem or in your VPC,
    wired to your data access.

  • A woman at a desk working on a laptop with code on the screen.

    Onboarding

    Your team trained on the agent,
    first runs on your own data.

  • A man wearing a headset working at his laptop in an open-plan office.

    Expert support

    Up to 20 hours of guidance from the Probabl team toward your first use case.

The only question left is whether it fits your data

The scope is fixed. The price is fixed. The deployment is in your cloud. What varies is whether Skore matches your team’s specific use case — and that’s a 30-minute conversation.

News on Skore, research on Tabular AI, and upcoming events from the Probabl team.