Hackathon Overview

There are two major types of hackathons on Pond: Model Competitions and Hackathons.

Model Competitions

Format: Kaggle-style challenges with specific datasets and scoring criteria

• Build a machine learning model to solve a specific task—such as Sybil detection or price prediction—in a Kaggle-style challenge. We partner with top-tier projects to host model competitions across diverse use cases. As a builder, you’ll have the opportunity to showcase your model, compete for rewards, and optionally deploy it on Pond to gain exposure and monetize through third-party usage.

How it Works

1. Browse active competitions
2. Download datasets or use Pond AI Studio
3. Build and iterate
4. Submit your models: You can submit at most 3 times per day
5. View the leaderboard: Many competitions provide scoring while the submission window is still open. Don't like your score? Submit another iteration and move up the leaderboard.
6. Deploy your model: Top performing models will be deployed to Pond Integration Hub for others to use and the model developers will share the revenue generated from the model usage.

Hackathons

Format: Project oriented under a general theme

• The goal is normally broader - you can build in different cases proposed by the hackathon hosts, such as building agents in finance applications, a calendar app with advanced voice synthesis technologies, or walk in the randomness if the host proposed a hackathon with a theme of the most creative application in general (in that case you can develop freely and the market, judges and hackathon organizers will choose the best ones). The outcome can be more free-form.
• Top-performing hackathon participants are eligible to launch their projects on Pond Markets, gaining access to our infrastructure and acceleration support. Recognized by both the community and the market, these projects continue their journey with backing from Pond and early users.