Serverless ML (@ServerlessML) header
Serverless ML (@ServerlessML) profile picture

Sharing insights into the future of Machine Learning through the lens of no-infra MLOps

Joined July 2021
Hi, Everyone! The countdown has started; this year's Feature Store Summit is just a couple of hours away. We hope you're as excited as we are! Don’t forget to join the event through Zoom: us02web.zoom.us/j/8506903657… See you soon!
15
🚀 Over 1,600 registrations. 13+ talks. Free. Online. Hear how Uber, Lyft, Pinterest, Zalando & more push production ML & Real-Time AI. You don't want to miss this: featurestoresummit.com/regis… No time that day? Register to get all the recordings.
18
Revolutionize your face recognition with this serverless #ML project! The pipeline automatically retrains to ensure optimal accuracy, giving you the best results every time. Say goodbye to manual updates and hello to seamless facial recognition bit.ly/3FGDHtP
2
1
96
Now you can predict the price of an apartment in Stockholm with @Hopsworks!  This #Stockholm Housing Valuation will tell you the best agencies to use and a stipulation of the home's value.  ➡️ bit.ly/3kCBzvS
2
5
301
Will your post on #Reddit be successful? 🤔 Check out this scalable machine learning system to predict the number of likes a Reddit post or comment is going to get if the user decides to post it 🚀 ➡️ bit.ly/3KFyVAf
3
4
421
Would you like to start investing in stocks? Is Tesla the way to go? This #ML project helps you. It uses time-series data for training an LSTM model, using @Hopsworks and Modal. ➡️bit.ly/3ZjVsa0
1
5
535
Us all day, everyday 💕
4
The Issue 32 of our Newsletter is out! We talked about the Feature Store Summit, more news about the Serverless course, the coolest events, and so much more. Subscribe now! 🔥featurestore.org/
All #FSS2022 are live! Watch Lu Mian from @4ParadigmHealth in his talk 'OpenMLDB: An Open-Source Real-Time Feature Platform Computing Consistent Features for Training and Inference.' He introduces OpenMLDB 🚀 📹bit.ly/3sdT31J
1
Today deserves a #FSS2022 talk! Enjoy 'ralf: Real-time, Accuracy Aware Feature Store Maintenance by @sarahwooders, Ph.D. @UCBerkeley. She presents the notion of feature store regret that helps evaluate the feature quality of different maintenance policies bit.ly/3SkB1ps