AI Hackathon
← DEPARTURES
Open ThemeBoarding

Frontend Arena · AI BuildFest 2026

Frontend Arena · on Unstop

Cash prize: $301
Free credits: Credits: not stated
Who can enter: Team required Country rules: unknown
Code rights: Not checked yet
AdaptionAdaptive DataInvent-a-DatasetAutoScientist
Registration closesDec 8

A 48-hour online sprint where teams train their own model on Adaption and ship a working solution to a challenge revealed at kickoff.

Proceed to gate Opens unstop.com. You register on the organizer's site.

Imported from Unstop and approved by our team. We haven't checked the rules by hand. Confirm the rules on the event page before you enter.

One email a week

New events that match your topics, every Monday. Unsubscribe anytime.

Flight details

Registration closes:
2026-12-08 18:29 UTC
Submissions close:
2026-12-12 06:30 UTC
Event ends:
2026-12-12 06:30 UTC
Cash prize:
$301
Page language:
English

Prizes listed: Winner: amount not stated (Overall Winner of AI BuildFest 2026. Recognition for the team presenting the strongest overall solution, based on the hackathon’s judging criteria. Any platform credits or additional benefits are subj); 1st Runner-up: amount not stated (First Runner-up of AI BuildFest 2026. Recognition for the team achieving the second-highest overall evaluation score. Any platform credits or additional benefits are subject to the final sponsor prize); 2nd Runner-up: amount not stated (Second Runner-up of AI BuildFest 2026. Recognition for the team achieving the third-highest overall evaluation score. Any platform credits or additional benefits are subject to the final sponsor prize); Adaption Leaderboard Prize: INR 19,000 (Approximate INR equivalent of the USD 200 cash prize sponsored by Adaption. Awarded to the participant who reaches #1 on any Adaption domain leaderboard at any point during the 48-hour hackathon, even); Best Use of Adaption: INR 9,500 (Cash prize equivalent to USD 100, sponsored by Adaption. Awarded to the team demonstrating the most effective and meaningful use of Adaption’s platform during AI BuildFest 2026. Prize amount is an app)

In short

  • Submissions are due 12 Dec 2026 at 06:30 UTC and registration closes 8 Dec 2026 at 18:29 UTC. Confirm both on the site.
  • ₹28,500 in cash: ₹19,000 for first place on an Adaption domain leaderboard and ₹9,500 for Best Use of Adaption.
  • Projects must build a dataset with Adaptive Data or Invent-a-Dataset, train with AutoScientist, and teams need 2 to 4 members.

Project ideas

Material Photo Sorter

Generate a labeled material-photo dataset, train a classifier in AutoScientist, and show predictions for uploaded photos.

Architecture: Collect material photos as raw examples → Generate labeled dataset in Invent-a-Dataset → Train classifier with AutoScientist → Show predicted material for an uploaded photo

Stack: Invent-a-Dataset, AutoScientist, Adaption platform, HTML/JS upload page

Held-Out Split Reviewer

Create a dataset, train a model on Adaption’s platform, and compare held-out predictions to find weak spots.

Architecture: Create labeled examples with Adaptive Data → Split examples into train and held-out test sets → Train model with AutoScientist on Adaption → Display test accuracy and error examples

Stack: Adaptive Data, AutoScientist, Adaption platform, evaluation dashboard

Adaptive Data Cleaner

Use Adaptive Data to clean and label examples, then train a model with AutoScientist and show predictions.

Architecture: Collect raw examples for a chosen task → Clean and label examples in Adaptive Data → Train model with AutoScientist → Show predictions for new examples

Stack: Adaptive Data, AutoScientist, Adaption platform, simple demo UI

FAQ

Who is eligible to enter?
The listing names eligible groups as "All" and describes the event as bringing together students, developers and AI enthusiasts, but eligibility is not confirmed. Check the official rules before registering.
Who owns the code and the trained model?
The event page does not say who owns the code; read the official rules. It does say participants keep their trained weights and may publish them on Hugging Face or Kaggle, subject to platform terms and data licensing.

Sources

Open the event page →

More open AI hackathons