Empirik has emerged from stealth with $21 million in funding to help enterprises modernize how they understand and act on infrastructure change. The company was initially built inside Sequoia by Avon Puri and Sudheer Dhurjati, both experienced infrastructure operators from Rubrik and VMware. It is now led by Chief Executive Officer Kartik Chandrayana, a repeat founder who previously held product leadership roles at Salesforce and Quantum Metric.
The Shortcomings of Traditional Observability
Traditional observability platforms generate metrics, logs, and traces that primarily describe failures after they occur. The real driver of infrastructure behavior is change, such as a new identity policy, a completed Terraform provider, or a feature flag flipped in production. Yet change has often been treated as an administrative or compliance task rather than a source of operational intelligence.
Why Infrastructure Has Lagged Behind
Agentic coding has advanced because code lives in version controlled repositories that make architecture and intent legible to machine learning models. Infrastructure has lacked this foundation, as infrastructure as code often reflects a declared state rather than what is actually running in production. As a result, there is no unified, real time representation spanning cloud resources, Kubernetes, networking, identity, and legacy systems.
Change as the Missing Pillar
Observability is already a proven market, with companies such as Datadog, Dynatrace, and Splunk generating billions of dollars in revenue. However, these platforms are built to measure symptoms like latency spikes, rising errors, or crashed pods. Empirik argues that change itself is the missing fourth pillar, telling teams that a specific deployment caused three virtual machines to shut down instead of merely reporting a latency increase.
An Infrastructure Compiler for Real Time Understanding
At the core of Empirik is an infrastructure compiler that converts metadata from clouds, identity providers, CI/CD pipelines, and SaaS systems into a live operational graph. This graph maps thousands of relationships between assets and accounts so teams can see what changed, who changed it, and whether it caused an incident. The company also connects directly to agentic workflows through MCP, allowing tools like Claude and ChatGPT to reason over infrastructure change as part of an Autonomous Infra Engineer.
Enterprise Traction and Early Results
Empirik is already processing millions of raw change and telemetry events each week across dependency graphs that span millions of resources. TCBPay, a payment processing company with more than $1.5 billion in annual volume, now catches sensitive configuration file changes in ten seconds instead of thirty minutes. Guardant Health expanded its deployment from production to all development environments, while a global Fortune 50 consumer packaged goods company is using Empirik to support much of its autonomous infrastructure roadmap.
Leadership Built for Enterprise Scale
Avon Puri and Sudheer Dhurjati designed the product from day one for the complex hybrid stacks common in Fortune 500 enterprises, spanning public clouds, on premises infrastructure, identity layers, and critical SaaS applications. CEO Kartik Chandrayana brings firsthand expertise in observability and big data from Salesforce, where he led product after selling his first company Twin Prime to the software giant. Their combined experience is aimed at answering the critical question of what changed and whether it caused an incident.
With $21 million in new funding, Empirik is positioning itself as the system of record for infrastructure change for humans today and for autonomous agents in the future. The company is moving beyond symptom based observability to provide causality and prevention before risky changes reach production. Its early deployments suggest that enterprises are seeing faster detection, reduced risk, and more complete operational context across hybrid environments.