Kaggle
In plain wordsA popular website, owned by Google, where people compete to build the most accurate prediction models on shared datasets, publish their code and data, and learn data science.
An online platform for data-science and machine-learning competitions, datasets, shared notebooks, pretrained models and courses. In a typical competition the host supplies training data and a hidden test set; participants submit predictions scored on a public leaderboardleaderboard, n.: A public ranking table showing which AI systems score best on a test, updated as new results arrive — like a sports league table for models. computed on part of the test set, while final standings use a private leaderboard on the rest, which punishes overfittingoverfitting, n.: When a model learns its training examples too well — quirks and noise included — so it does great on data it has seen but poorly on anything new, like a student who memorizes answers without understanding. to the public scores. A progression system ranks users from Novice through Contributor, Expert and Master to Grandmaster in competitions, datasets, notebooks and discussion. Kaggle shaped practical machine learning: competitions popularized XGBoostXGBoost, prop. n.: A fast, free software tool that makes predictions by adding up hundreds of small decision trees, each one fixing the mistakes of those before it. For years it was behind many winning entries in competitions on spreadsheet-style data. and later LightGBMLightGBM, prop. n.: A widely used tool for making predictions from very large tables of data, quickly and in little memory, by combining many small decision trees. Microsoft released it as free software in 2016. and CatBoostCatBoost, prop. n.: A free gradient-boosting library from the tech company Yandex, known for handling categories such as city names or product types directly, without first converting them into numbers by hand., heavy feature engineeringfeature engineering, n.: Turning raw data into the inputs a model learns from, using knowledge of the problem; for example, converting a timestamp into 'hour of day' and 'is it a weekend'., careful cross-validationcross-validation, n.: A fair way to test a model when data are limited: split the data into several parts, train on all but one, test on the part left out, and rotate so every part takes a turn. and stackedstacking, n.: Combining several different models by training one more model to decide how much to trust each of them, based on how they performed on data they had not seen. ensembles, and its pragmatic, leaderboard-driven craft became known as Kaggle culture. It also hosts research competitions such as the ARC Prize and the AI Mathematical Olympiad progress prizes, and as of April 2025 reported over 23 million accounts.
The Kaggle winners' write-up shows how much of the gain came from validation strategy, not the model.
Origin
Founded in April 2010 by Anthony Goldbloom; Ben Hamner, its longtime chief technology officer, is also counted among its founders, and Jeremy Howard joined as president and chief scientist that November. Early contests such as the Heritage Health Prize and the 2014 Higgs Boson Machine Learning Challenge made 'Kaggle' shorthand for competitive machine learning. Google's acquisition was announced on 8 March 2017; Goldbloom and Hamner left in 2022, when D. Sculley became chief executive.
Usage note
Critics note that competition wins can reward leaderboard tuning and leakage over deployable models. In 2025–2026, retractions of medical papers built on poorly sourced Kaggle datasets drew attention to data provenance and consent on the platform.
Related
- XGBoostXGBoost, prop. n.: A fast, free software tool that makes predictions by adding up hundreds of small decision trees, each one fixing the mistakes of those before it. For years it was behind many winning entries in competitions on spreadsheet-style data.
- LightGBMLightGBM, prop. n.: A widely used tool for making predictions from very large tables of data, quickly and in little memory, by combining many small decision trees. Microsoft released it as free software in 2016.
- stackingstacking, n.: Combining several different models by training one more model to decide how much to trust each of them, based on how they performed on data they had not seen.
- leaderboardleaderboard, n.: A public ranking table showing which AI systems score best on a test, updated as new results arrive — like a sports league table for models.
- tabular datatabular data, n.: Data arranged as a table, like a spreadsheet: each row is a case — a customer, a loan, a patient — and each column a measured property. Much business and scientific data looks like this.
- data leakagedata leakage, n.: A mistake in building a model that lets it peek at clues it will not have in real use, such as hints about the very answer it must predict, so it shines in testing and then disappoints.
- benchmarkbenchmark, n.: A fixed, shared test — a set of questions or tasks with an agreed way of scoring — used to compare AI systems with one another and to track progress over time.
Mentioned in
ARC-AGIARC-AGI, prop. n.: Puzzles made of colored grids — and, since 2026, small video-game-like worlds — that people solve easily but AI long found hard, designed to test whether a machine can learn a new skill from just a few examples., leaderboardleaderboard, n.: A public ranking table showing which AI systems score best on a test, updated as new results arrive — like a sports league table for models., data sciencedata science, n.: The practice of collecting, cleaning, analyzing and interpreting data, using statistics, programming and machine learning, to answer questions and guide decisions., Jupyter notebookJupyter notebook, n.: An interactive document, opened in a web browser, that mixes runnable code with its results, charts and notes, so people can experiment step by step and share their work., LightGBMLightGBM, prop. n.: A widely used tool for making predictions from very large tables of data, quickly and in little memory, by combining many small decision trees. Microsoft released it as free software in 2016., tabular datatabular data, n.: Data arranged as a table, like a spreadsheet: each row is a case — a customer, a loan, a patient — and each column a measured property. Much business and scientific data looks like this., XGBoostXGBoost, prop. n.: A fast, free software tool that makes predictions by adding up hundreds of small decision trees, each one fixing the mistakes of those before it. For years it was behind many winning entries in competitions on spreadsheet-style data., MLE-benchMLE-bench, prop. n.: A test that gives AI agents 75 real data-science competitions from Kaggle and checks whether their solutions would have won medals against human teams., stackingstacking, n.: Combining several different models by training one more model to decide how much to trust each of them, based on how they performed on data they had not seen.