Case Study: Nooks achieves 250ms median dialer latency with Porter

Nooks achieves 250ms median latency and cuts infra costs 50% with Porter

Nooks, an AI sales technology startup, faced significant infrastructure challenges that impacted their core product, an AI parallel dialer. They initially used a multi-vendor approach, including Google Cloud Functions and Fly.io, but encountered problems such as cold starts causing call connection delays and execution timeouts that hampered data imports. These issues, along with a lack of control and visibility on their PaaS, threatened the user experience for their sales rep customers and limited their ability to scale.

By migrating to Porter, Nooks consolidated its infrastructure onto Kubernetes within its own private cloud. Porter provided the managed ease of a PaaS while giving Nooks the control to eliminate cold starts, configure critical application parameters, and deploy GPU-accelerated inference workloads. The solution delivered a median call connection latency of 250ms, nearly 50% cost savings on infrastructure, and eliminated the risk of vendor lock-in, allowing the small engineering team to focus entirely on product development.

Nooks