Empirik wants to retire the emergency page. The company, incubated inside Sequoia Capital and now independent, has raised $21M in seed funding from Sequoia, Canapi, and Alumni Ventures to predict infrastructure outages before they happen.
The founding pair came from inside Sequoia’s own operations. Avon Puri, the firm’s chief digital and information officer with infrastructure stints at Rubrik and VMware behind him, started the project with colleague Sudheer Dhurjati. Their thesis: language models could catch failure patterns that escape human engineers. Kartik Chandrayana, an ex-Quantum Metric CPO and former Salesforce observability VP, was later brought in to run it.
The product watches every change engineers make across sprawling systems and maps the likely fallout. Low-risk updates pass automatically. Larger ones get guardrails. The riskiest changes route to human reviewers. For DevOps and site reliability teams, the pitch is less time chasing incidents and more time on work that matters.
Sequoia partner Bogomil Balkansky argues traditional observability tools never grasped system dependencies, leaving Empirik alone in its category, layered above AI site reliability platforms like Resolve and Sequoia-backed Traversal.
Customers already span startups and Fortune 500 names, including Guardant Health and a major consumer packaged goods company. Chandrayana borrows the developer-tools analogy: Cursor and Claude Code automated software work, and Empirik aims to do the same for the engineers who keep systems running.
