@SachitBansal20@Saumilya_gupta@JuniorVasu Yeah same with my team we had a rank of 5 and bow arent even in the top 50 , because of the candidate pairs but should have mentioned expliclity a very shitty shakeup this time
We took a break at rank 30. We came back and found ourselves at rank 126.
That was the lowest point of the Amazon ML Challenge 2026
72 hours, 89K+ registrations, and one brutal problem. Match 1.7M businesses against ~10M messy records across the US, India and France.
Here's what nobody tells you about a live leaderboard.
it never stops moving. Stand still for 30 minutes and you lose a rank or two. So you don't get to stand still.
Our journey, rank by rank:
8 → 12 → 30 → 126 → 36 → 16 → 8 → 5
Looking at that sequence now, the 126 is the part I'm proudest of. It's where we could have quietly decided the game was over.
Instead we kept debugging, kept shipping, and climbed back.
Final score: 0.99153.
The last 0.001 took more work than the first 0.05. That's the part no leaderboard screenshot shows.
Rank 5 out of 89K+. Grateful to @Arushhh_Sharma , who stayed in the trenches with me through every one of those ranks.
#AmazonMLChallenge #MachineLearning #AmazonML2026 #IITBombay @AmazonScience
@_AnupamRawat_ Thanks, @_AnupamRawat_ . Rank 5 out of 89K+ registrations, on a leaderboard that reshuffled every 30 minutes. Entity resolution at 10M-record scale is where I want to keep working.
Rank 5 on the public leaderboard, Amazon ML Challenge 2026 🏆
72 hours, 89K+ participants and a leaderboard that moved every half hour: 8 → 12 → 30 → 126 → 36 → 16 → 8 → 5.
Final score : 0.99153
Thanks @_AnupamRawat_ for the compute and company.