We built Kubestead because we were tired of writing postmortems.

After years running platform infrastructure at growing technology companies, our founders kept encountering the same failure mode: a canary passes the readiness probe, reaches 60% traffic, and breaks production two hours later — because the metric that mattered was never in the dashboard anyone was watching. Kubestead automates the decision loop so the right metric is always in the loop.

Engineering team reviewing a deployment dashboard on shared monitors
Thomas Rivera, CEO and co-founder
Founders

Thomas Rivera, CEO & Co-founder

Thomas spent eight years in SRE and platform engineering before starting Kubestead. In his last role before founding — Head of Platform at a growing B2B technology company — he oversaw infrastructure for 200+ microservices and personally led rollback procedures for 14 production incidents in a single year. Eleven of those incidents shared an identical root cause: a metric threshold that wasn't being watched during the canary window. In 2023, he moved back to Miami and co-founded Kubestead to build the automated analysis loop he'd been assembling by hand. "Every postmortem said the same thing: we should have caught this in canary. We didn't because it took three people, two dashboards, and a gut-feel call to make the right decision. Kubestead makes that call automatically — without the gut-feel."

miami, fl founded 2023 angel: $2m
Team

The people building it

Thomas Rivera

Thomas Rivera

CEO & Co-founder

8 years SRE and platform engineering. Managed 200+ microservices and ran rollback procedures for 14 production incidents in one year. Founded Kubestead in Miami in 2023.

Lucia Ferreira

Lucia Ferreira

CTO & Co-founder

Kubernetes contributor since 2020. Spent 6 years building observability pipelines at two cloud infrastructure companies. Wrote the first internal progressive delivery tooling Kubestead is built on.

Raj Mehta

Raj Mehta

Head of Engineering

Kubernetes Operator pattern specialist. Former tech lead at a developer tooling company serving 10,000+ engineers. Obsessed with mean-time-to-recovery and the metrics that predict it.

Sofia Andersen

Sofia Andersen

Head of Growth

5 years in developer tools product and growth. Previously launched onboarding for a GitOps platform adopted by 3,000+ teams. Believes the fastest GTM for infrastructure tools is a 10-minute quickstart and honest documentation.

How we build

01

Least privilege, always

We hold ourselves to the same standard we ask of production infrastructure. Kubestead's internal services operate with minimal IAM permissions, automated secret rotation, and per-service mutual TLS. The controller ClusterRole is scoped to exactly what it needs — no more. Our security posture is documented, public, and auditable.

02

Remove humans from the failure path

Safe progressive delivery should not require anyone to be watching a dashboard at 2 AM. Every feature we ship is evaluated against a single question: does this reduce the decisions an on-call engineer has to make, or does it add to them? We do not build features that require human interpretation to be safe.

03

Open where it counts

The controller is Apache 2.0 open source. The RolloutPolicy CRD schema is fully documented. Security advisories are public on GitHub. You do not have to trust our marketing — you can read the code, audit the RBAC model, and validate the rollback logic before you give Kubestead access to your cluster.

Join us in building it.

We're a small team in Miami building deployment infrastructure for SRE teams. If you care about observable, predictable rollouts, we want to talk. Miami-based or remote.