
Catch faults before failure
Detect early warning signs in sensor data to help maintenance teams prevent breakdowns and costly downtime.
AI learned to see with shared image data. TimeNet brings time-series data together to help AI understand how the world changes.
Video unavailable. Open the video.
Pick a time-series problem, onboard a dataset in TimeNet format, then train or fine-tune a Temporal AI Model (e.g. OpenTSLM) to explore what it can do.
Choose any use case, from forecasting and classification to explanation, reasoning and beyond.

Detect early warning signs in sensor data to help maintenance teams prevent breakdowns and costly downtime.

Spot inefficiencies in operational data to help operators reduce energy losses and get more from each turbine.

Train a model to classify ECG patterns and help researchers better understand unusual heart rhythms.
Have another idea? Build in any domain.
Pick a use case, consider who it could help and find a relevant time-series dataset, or bring your own.
Onboard your dataset in TimeNet format, with the signals, metadata and annotations your task needs.
Build for your chosen task, evaluate on held-out data and share your results.
Share a demo, pitch, website or another format. Explain what works, its limitations and how someone could benefit from it.
Submissions close at 11:59 p.m. PST (UTC−8). Judging takes place the following Monday.
Two weeks to build. More rounds to come.
$100 or $500
A Nebius voucher for your team, one per team, per round. Applications are reviewed on a rolling basis, based on your idea’s potential and credit availability.
Starting at$25,000*
For winning teams.
* T&C apply.
Starting at$25,000*
For winning teams.
* T&C apply.
Round winners and overall winner
More rounds and prizes to come, and teams can win more than once.
Register and join Discord. Start with GitHub and the quickstart.
Registration is required for everyone.
Nebius: apply through the registration form with your project idea for a $100 or $500 voucher, one per team, per round. There is no fixed application deadline. Vouchers are allocated on a rolling basis, based on project potential and availability.
AWS: starting at $25,000 in credits* for winning teams.
Google Cloud: starting at $25,000 in credits* for winning teams.
* T&C apply.
Everyone can register and participate, with or without credits. Credits are not guaranteed. We’ll share setup guidance in Discord.
Round 1 submissions close November 22 at 11:59 p.m. PST (UTC−8). Judging takes place the following Monday.
More rounds are coming, and we’ll announce them in Discord. Teams can enter more than one round and submit a different use case each time. Each round has its own winners, plus one overall winner across all rounds.
Each round has its own winners and prizes, plus one overall winner across all rounds. Teams can win more than once.
Submit one entry per team for each round you enter:
The clarity of your public X and LinkedIn posts and their traction are part of judging. Share your progress early and include screenshots of available metrics with your submission.
Get help in OpenTime on Discord during European business hours.