Platform engineering

Your product needs to deploy to Kubernetes, scale across cloud providers, and ship updates multiple times per day without breaking production. We’ve built this infrastructure for AI products, developer tools, and security platforms. We bring open source tooling we created specifically for this work.

What we’ve built

We’ve been StackBlitz’s infrastructure partner since 2021—building the backend and performance systems that support 3M+ monthly developers and bolt.new’s explosive growth to $40M+ ARR. For Whop, we cut their Rails test suite and CI time in half and executed a zero-downtime migration from PostgreSQL to PlanetScale.

Teleport is a privileged access management platform used by 100+ enterprise customers. We built their Go-based plugin ecosystem and deployment infrastructure. NATS is an open source messaging system for distributed systems. We improved their Ruby SDK performance by 3x. BetterUp needed a scalable AWS backend for their AI coaching platform ($570M+ in funding). We built it.

Kubernetes and deployment

Our Martian Kubernetes Kit is a battle-tested toolkit for running production workloads on Kubernetes—covering deployment, monitoring, autoscaling, and security out of the box. We use it with our clients and ship it as open source.

For teams not ready for Kubernetes, we work with Kamal for zero-downtime deployments, Docker-based development environments, and managed cloud platforms on AWS and GCP.

Multi-tenancy and self-hosted deployments

Enterprise customers want your product on their infrastructure. That means multi-tenant isolation, per-tenant resource limits, data residency compliance, and deployment automation that works across AWS, GCP, and on-prem Kubernetes clusters. We’ve built this for bolt.new—on-prem Kubernetes deployments for enterprise customers—and we bring that experience to every product that needs to deploy anywhere without the multi-tenancy pain.

We handle fair job prioritization so one tenant’s bulk import doesn’t starve the rest, and the deployment tooling to ship updates across all environments simultaneously.

CI/CD and developer productivity

Slow CI kills momentum. We profile test suites with TestProf (our open source tool), eliminate flaky tests, optimize build pipelines, and set up monitoring that tells you what’s wrong before users notice. The goal: your team deploys confidently, multiple times per day—with alerting, distributed tracing, and infrastructure as code via Terraform.