Cardinal is an observability data lake and agent runtime. We give AI agents complete evidence, machine-scale investigations, and infrastructure that runs entirely inside your cloud.
Our Story
We spent years at Netflix building the platforms that powered the company's engineering teams — petabyte-scale observability, real-time data pipelines, and the developer tools used every day by thousands of engineers to keep hundreds of millions of members online.
We saw firsthand what great infrastructure lets teams accomplish — and what happens when the tooling underneath falls behind the workload on top of it.
Today, the operators of production aren't just humans. Agents are opening pull requests, debugging incidents, and running investigations at a scale no dashboard was ever built for. Cardinal is the layer we always wished we had — one that lets agents reason over the full haystack without ever compromising on control, cost, or privacy.
“The next wave of software won't be shaped by anyone tuning models. It'll be shaped by engineers building agents that solve real problems — and the tooling underneath them is the bottleneck.”
Where we came from
7 yrs
leading Netflix's observability platform
Petabytes
of daily telemetry processed
1000s
of engineers served every day
100Ms
of Netflix members kept online
What we believe
Full fidelity, always
Sampling is a symptom of bad economics.
Every downsampling decision throws away the exact detail an investigation later needs. We designed Cardinal so full-fidelity telemetry is the default, not a premium tier.
Your cloud, your rules
Telemetry never leaves your VPC.
Cardinal deploys inside your account, writes to your object storage, and gives each investigation its own compute. No exfiltration, no shared multi-tenant surprises, no vendor lock-in on your data.
Investigations become software
Every successful run compiles into a reusable skill.
Discovery is expensive. Repetition should be cheap. Cardinal turns each successful investigation into deterministic mechanics — bounded, repeatable, and available to the next agent that needs it.
Meet the founding team
Ruchir Jha
CEO
Ruchir led Netflix's Observability Platform Team for 7 years, architecting the petabyte-scale monitoring and alerting that kept Netflix running for hundreds of millions of members. He started Cardinal to build the infrastructure layer the next generation of agentic engineering will run on.
Michael brings deep systems and open-source experience — from co-authoring IETF standards work on DNS (RFC 6891) to contributing to core internet infrastructure like BIND and deployment-safety tooling in the Spinnaker / Kayenta ecosystem.
