Cerebrium's just launched on Product Hunt! We're a serverless cloud infrastructure platform built specifically for ML apps. After experiencing the pain of managing ML infrastructure firsthand, we built what we wished existed: Key platform features: • 2-4s cold starts (vs industry standard 30s+) • Speedy build and iteration times (We've seen subsequent builds take under 9s to build and deploy to live) • 8+ GPU types (H100, A100, L40s) • Native streaming/websocket support • ASGI app support • Multi-GPU capabilities • No complicated syntax and a short learning curve for getting up and running on the platform • Just deploy your Python code We're currently supporting ML teams from seed to Series C, and we've open-sourced several implementations including: • OpenAI Voice alternative (faster & cheaper) • Real-time AI avatar system • Live stream product detection Try it with $30 free credits: dashboard.cerebrium.ai If you experience value on the the platform, consider upvoting, commenting or leaving us a review on Product Hunt: https://www.producthunt.com/posts/cerebrium