Lessons from Airbnb’s AI experiments, including what worked, what failed, and why simplicity often wins.
Loved the breakdown Jose!
Building a NN that scales is for sure not an easy feat.
I’ve been considering it as a replacement for a Gradient Boosted Tree model I’ve been building at work for the past 5 months but which has reached its natural ceiling.
Lots to consider here…
Thanks Andres!
I have been building NN for some time and there are lots of quirks that can help.
The tree based mindset needs to change both for feature engineering, for hyper parameters tuning and also for deployment.
Happy to help if you need tips!
Loved the breakdown Jose!
Building a NN that scales is for sure not an easy feat.
I’ve been considering it as a replacement for a Gradient Boosted Tree model I’ve been building at work for the past 5 months but which has reached its natural ceiling.
Lots to consider here…
Thanks Andres!
I have been building NN for some time and there are lots of quirks that can help.
The tree based mindset needs to change both for feature engineering, for hyper parameters tuning and also for deployment.
Happy to help if you need tips!