← Journal · · Announcement

Part one of the article about the inference platform is out!

Hey everyone :) We published part 1 of the article based on my HighLoad++ talk 🕺🕺🕺

In it I laid out the platform requirements, how we started with Seldon and why we moved away from it, and the current shape of our platform!

So it's a great time to start the work week by reading the new article :)

In the next part there'll already be hardcore technical meat, so don't worry :) I'll tell you how we implemented resource autoscaling, canary deploy, the inference graph, and how you can automate picking a Triton configuration. Part 2 is coming soon, so stay tuned to the channel!

Original on Telegram ↗

↑↓ select · Enter open · Esc close