If We're Not Up, The Checkout Breaks

Kintsugi puts sales tax compliance on autopilot for 7,300 companies selling into 110 countries, and its tax engine sits inside other companies' checkouts. The answer has to arrive before the shopper finishes paying. "Typically, a customer's expectation is that we are returning the sales tax estimate on the invoice within 100 milliseconds." The cost of missing is not abstract: "For every minute that our sales tax API, if it ever goes down, our customers are not able to collect roughly $4 million in sales tax that they should be collecting."

Co-founder and CEO Pujun Bhatnagar and Engineering Lead Marcus Bertrand walk through what that pressure did to their first architecture. Kintsugi started on AWS Lambda, and the limits arrived with the customers: memory and CPU bound together, hot workers pre-provisioned or requests timing out, and concurrency ceilings that failed silently. "We would hit these limits and just kind of silently fail, and then discover that a little bit later."

So they decided almost immediately not to go full native with any one cloud provider. On Control Plane they kept their database and SQS in AWS, had a working proof of concept within hours, and migrated piece by piece with a sixty-second rollback whenever a test went wrong. Today the API workloads scale on latency through daily traffic curves and customers calling 3,000 times a second, and the infrastructure conversation inside Kintsugi is mostly over: "We don't worry about the infrastructure. We're worried about delivering product."

Chapters 0:00 Sales tax on autopilot 0:37 Salt in a dish: inside the checkout 1:05 Under a second, or the tax isn't collected 1:35 $4 million a minute 1:47 Two founders to 7,300 customers 2:37 The daily curve, and scaling on latency 3:35 Silent failures on AWS Lambda 4:06 Multi-cloud, not native 4:38 A migration with a sixty-second undo 6:03 Data where customers need it 6:47 Agents, sandboxes, and what's next 7:45 "We don't worry about the infrastructure"

Read the full Kintsugi case study: https://controlplane.com/customers/kintsugi

Control Plane combines AWS, GCP, Azure and your own hardware into one virtual cloud shaped to your workloads. Run a workload in any cloud or region, keep identity and networking consistent across all of them, and pay only while workloads are running.

Start free: https://controlplane.com Docs: https://docs.controlplane.com