Introducing Cleric: The first autonomous AI site reliability engineer
We’re excited to announce that we’re building an autonomous AI SRE, called Cleric, backed by Zetta Venture Partners in a $4.3M seed round. Cleric is an AI teammate designed to autonomously manage, optimize, and heal software infrastructure.
We're building an AI SRE
Production environments suck. We’ve learnt this the hard way. Dinners interrupted by alerts. Explaining lost revenue to leadership. We’ve even faced a rioting mob due to a simple misconfiguration (ask us about this!).
While others are focused on applying AI to software development, our mission is to free up engineers from production toil. Cleric is an AI agent that will autonomously manage, optimize, and heal your software infrastructure. The holy grail.
Our industry is obsessed with building more tools for engineers. The problem with tools is that they require human operators. We don’t scale out like machines, we can’t process large quantities of information, and we make mistakes. Large-scale production environments simply aren’t meant for humans. We need closed-loop infrastructure, where machines operate machines. Extending existing tools with AI won’t get us there.
We’re building an operator, not another tool.
The teams that adapt to this new AI paradigm will unlock orders of magnitude more productivity. Agentic infrastructure is an ambitious project, so we’re taking an incremental approach.
We're starting with an on-call teammate for software engineers
The first concrete step for Cleric is to free on-call engineers from time consuming investigations. Today, Cleric is able to reliably triage and root cause alerts from production applications. We’ve trained Cleric extensively on complex Kubernetes based scenarios, but it is built from the ground up to adapt to idiosyncratic enterprise environments.
Cleric’s ability to reason lets it handle a very wide range of issues, and at speeds much greater than that of manually stepping through a runbook. It generates hypotheses using your teams’ context, collects evidence by running queries for configuration, logs, or metrics, then processes these findings to diagnose problems just like an engineer would.
Cleric is safe, extensible and powerful. Today, it only has read access to your infrastructure and will make suggestions, but not apply them. All integrations are source-available and extensible. You can even teach Cleric how to integrate with your internal APIs. Cleric does not sleep and can handle thousands of issues concurrently.
Cleric is now in early access for software engineering teams with on-call responsibilities. If you’re drowning in on-call support, come and chat with us.
We're builders with experience in infrastructure and AI
Shahram was the Head of Engineering for MLOps, container deployment, and FinOps at Gojek, the ride-hailing Decacorn in Southeast Asia.
Willem is best known as the creator of Feast, the widely adopted open source feature store. He led open source efforts at Tecton and founded Gojek's ML Platform team.
Together we’ve secured $4.3M in seed funding led by Zetta Venture Partners, with participation from AI infrastructure angel investors including leaders at Google Cloud, Sysdig, Tecton and Neo4J.
Want to help us build Cleric? We’re hiring in San Francisco and Singapore
Want to try Cleric? Get in touch!