Empirik.ai has emerged from stealth with a $21 million seed round backed by Sequoia Capital, S32, Canapi Ventures and Alumni Ventures, as the company develops an autonomous infrastructure engineer designed to help enterprises safely manage infrastructure changes at machine speed.
The company is led by CEO Kartik Chandrayana, who previously ran more than 30 infrastructure products at Salesforce. Chandrayana said the rapid adoption of AI coding agents is creating a growing mismatch between software development, which can increasingly move at machine speed, and infrastructure operations, where changes still often require manual human review.
Empirik.ai is designed to address that bottleneck by understanding the intent behind infrastructure changes, calculating their potential blast radius and determining whether changes can be safely executed autonomously or should be escalated for human review.
The platform is built around two architectural concepts.
First, Empirik treats change events as a primary signal before execution, enabling the system to evaluate infrastructure modifications before they are implemented rather than relying solely on observability after something has gone wrong.
Second, the company has developed what it describes as an infrastructure compiler that constructs a live operational model spanning on-premises infrastructure, public clouds, and SaaS applications.
The system also treats virtual infrastructure components such as IAM policies, access control lists and routing tables as first-class infrastructure objects, allowing the platform to map relationships and dependencies that could affect the impact of a proposed change.
Empirik.ai was created by Avon Puri and Sudheer Dhurjati and incubated within Sequoia Capital with support from partner Bogomil Balkansky. Chandrayana joined the company as CEO earlier in 2026.
The company says its technology is already running in production environments across Guardant Health, Avahi, TCBPay and multiple Fortune 500 companies.
Empirik’s platform is designed to operate across all major public clouds, on-premises environments and multiple SaaS applications, giving enterprises a unified model of infrastructure dependencies across increasingly complex technology environments.
The broader opportunity is being driven by the rapid expansion of AI-assisted software development. As coding agents increase the speed at which applications and infrastructure changes can be generated, Empirik believes enterprises will increasingly require an equivalent layer of intelligence capable of evaluating and executing infrastructure changes without making manual review the limiting factor.
KEY QUOTE:
“Running 30+ infrastructure products at Salesforce taught me one thing: speed and reliability were always at war. Now, that dynamic is breaking. Coding agents are shipping at machine speed, but infrastructure is still governed by human-speed manual review. You cannot put autonomous software development upstream of manually operated infrastructure and expect the system to hold.”
“Together, we set out to build the autonomous infrastructure engineer. We do this by capturing intent, computing blast radius, and safely evaluating whether a change can execute autonomously or needs human review.”
Kartik Chandrayana, CEO of Empirik.ai
