ML competitions and datathons (2026)
In an ML competition (also called a datathon) everyone gets the same dataset and problem. You train a model, upload predictions, and a live leaderboard scores them on a metric such as accuracy or macro F1. The best scores win.
Competitions on Hack4Hack run Kaggle-style: daily submission limits, a public leaderboard, and often a second round with a new dataset for the top teams.
Open and upcoming
ML Challenge Hack4Hack 2026
A two-round machine-learning challenge. A 24-hour Round 1 starts when the problem statement and dataset go live on 22 Oct; the top 100 get a new problem and dataset for 3 days.
Datathon by Hack4Hack · Online · ₹1L in prizes · Free · Registration closes 20 Oct 2026
Frequently asked questions
What is a datathon?
A datathon is a data-focused hackathon: teams analyse or model a dataset to answer a question or make predictions, and are ranked on a leaderboard or by judges.
How is an ML competition scored?
Your predictions on a hidden test set are compared with the true labels using the metric named on the hackathon page (for example macro F1). The leaderboard shows your best score.
Datathon vs hackathon: what is the difference?
A hackathon judges a working product and its demo. A datathon or ML competition judges the quality of your model or analysis on a fixed dataset. Read more in what is a hackathon.