Feature Store Evolution Under Cost Constraints: When Cost is Part of the Architecture by David Malinge & Ivan Burmistrov
ScyllaDB
90 views
0 slides
Mar 07, 2025
Slide 1 of 0
About This Presentation
At P99 CONF 23, ShareChat tackled scaling its ML Feature Store to 1B features/sec—then had to cut costs while maintaining SLAs. Ivan & David share how they optimized compute, reduced waste in Kubernetes, and tackled autoscaling for Apache Flink. It's geared for anyone interested in ML Feat...
At P99 CONF 23, ShareChat tackled scaling its ML Feature Store to 1B features/sec—then had to cut costs while maintaining SLAs. Ivan & David share how they optimized compute, reduced waste in Kubernetes, and tackled autoscaling for Apache Flink. It's geared for anyone interested in ML Feature Store and/or cloud cost optimizations.