For decades, enterprise infrastructure operations have remained trapped in a reactive posture. Despite organizations investing tens of billions of dollars into advanced observability tools—relying heavily on metrics, logs, and traces to signal when a system is failing—traditional monitoring has proven to be an inherently post-mortem exercise. It tells engineering teams that a service is down or lagging only after the damage has already been done.
Today, a well-funded startup named Empirik is stepping out of stealth to challenge that paradigm. Backed by $21 million in funding, the company aims to introduce true autonomy to enterprise infrastructure by shifting the industry from a reactive firefighting stance to a proactive operational model. Empirik was originally conceived and built inside Sequoia by seasoned infrastructure veterans Avon Puri and Sudheer Dhurjati, before spinning out as an independent entity under the leadership of CEO Kartik Chandrayana.
The core thesis behind Empirik stems from a familiar frustration shared by its founders. Before launching the company, Puri spent two decades running infrastructure at scale as the Senior Vice President and Chief Information Officer at Rubrik, as well as serving as Vice President of Engineering and Applications at VMware. Alongside Dhurjati, who worked side-by-side with him across three separate companies, Puri watched the same narrative repeat itself endlessly across large enterprises. Organizations were constantly playing catch-up with their systems, despite throwing massive budgets at observability platforms designed to diagnose symptoms rather than prevent root causes.
According to the founders, the true driver of infrastructure behavior has never been a system failure; it is change. Whether it involves attaching a new identity policy, completing a new Terraform provider deployment, or flipping a feature flag in a live production environment, change dictates how an environment evolves. Yet, historically, enterprise infrastructure has treated change as an administrative constraint. It has been relegated to static ticketing systems, manual approval forms, and tribal knowledge scattered across email threads and Slack channels. While changes are occasionally logged for regulatory compliance, they have rarely been utilized to execute actual infrastructure engineering tasks.
Empirik seeks to change that by capturing the real-time stream of changes across every layer of a modern technical environment. When an organization understands the continuous flow of changes happening throughout its architecture, it unlocks true operational autonomy. Every single change reflects an engineer’s intent. Once a system gains the capability to interpret that intent and accurately assess its potential impact, it can autonomously decide whether to execute the change safely or loop in a human operator. By bringing agentic capabilities to the infrastructure lifecycle, Empirik is addressing a glaring gap in enterprise modernization.
Why Change, and Why Now
While agentic coding tools have seen explosive adoption in recent years, infrastructure has largely lagged behind this technological wave. Coding tools took off because software code natively resides in version-controlled repositories, making an application’s underlying architecture and intent entirely legible to machine learning models. Infrastructure, however, has never possessed that standardized foundation.
Infrastructure-as-code represents a step forward, but because it typically outlines a declared state rather than mirroring what is actually running in live production, it remains fundamentally reactive. Enterprises lack a unified, real-time representation of their entire infrastructure across the sprawling maze of public cloud resources, Kubernetes clusters, networking layers, identity providers, and legacy systems. Unlike modern software engineering workflows, enterprise infrastructure largely lacks the concept of a pull request. Instead, every modification continues to run through the institutional memory of human operators rather than the rigorous logic of software.
Observability is already a massive, proven market dominated by multi-billion-dollar software giants like Datadog, Dynatrace, and Splunk. However, these platforms are purposefully built to measure surface-level symptoms—such as a sudden spike in latency, climbing error rates, or a crashing server pod. Empirik’s foundational bet is that there is a fourth pillar of observability hiding in plain sight: the changes themselves.
Instead of alerting a team that latency in a specific service increased by 300 percent over a 24-hour period, Empirik is designed to pinpoint the exact root cause, identifying precisely which configuration or deployment triggered the anomaly. Furthermore, the platform aims to intercept risks before they materialize, warning or blocking engineers at the inception of a change rather than scrambling to clean up the aftermath.
Building the Infrastructure Compiler
According to CEO Kartik Chandrayana—a repeat entrepreneur who previously founded Twin Prime before selling it to Salesforce, served as Salesforce’s Vice President of Product for Observability and Big Data, and most recently worked as the Chief Product Officer at Quantum Metric—infrastructure today mirrors the state where software engineering stood just three years ago. Before the advent of advanced developer coding assistants, every single action required a human in the loop because no reliable, comprehensive map of the system existed for machines to reason over.
Empirik is building that missing map, deploying it as an Autonomous Infrastructure Engineer designed to safely plan, reason about, and operate complex enterprise environments. The platform is already integrated into agentic workflows via the Model Context Protocol, connecting directly into developer tools like Claude and ChatGPT. This integration allows both human engineers and AI agents to evaluate infrastructure modifications using real-time operational context rather than guesswork.
At the core of the technology is a proprietary infrastructure compiler. Much like a traditional compiler translates raw source code into executable software, Empirik’s compiler converts metadata drawn from a customer’s multi-cloud environments, identity providers, continuous integration pipelines, and software-as-a-service applications into a live, unified operational graph. This graph maps thousands of intricate relationships across enterprise assets and accounts, giving the system real-time awareness of what changed, who implemented the change, and whether it directly contributed to a system failure.
Rather than taking shortcuts by building exclusively for modern greenfield Kubernetes environments or single public cloud providers, the founding team designed Empirik to handle the messy reality of Fortune 500 enterprises. Most large organizations operate complex, multi-decade hybrid stacks that span public clouds, legacy on-premise infrastructure, identity management layers, and critical SaaS applications. By constructing an operational graph that spans the entire enterprise surface area, Empirik ensures that both human teams and autonomous agents maintain complete contextual awareness.
Proving the Thesis in the Enterprise
The platform’s early commercial traction validates the market demand for a change-centric approach to infrastructure operations. Enterprise interest often traces back to costly outages stemming from unapproved or poorly understood modifications.
For instance, payment processing firm TCBPay, which handles over $1.5 billion in annual transaction volume, now catches sensitive configuration file changes in just 10 seconds, a dramatic reduction from their previous 30-minute detection window. Managed service provider Avahi integrated Empirik directly into their GitHub Actions workflows to evaluate and automatically block risky changes at the source. Meanwhile, organizations such as Guardant Health expanded their initial deployment from production environments to all development stages and on-premise systems after recognizing the platform’s utility. A global Fortune 50 consumer packaged goods company is similarly scaling its footprint, relying on Empirik to anchor its autonomous infrastructure roadmap.
Today, Empirik processes millions of raw change and telemetry events every week across dependency graphs that span millions of resources across major public clouds and legacy on-premise systems. Customers report measurable improvements across multiple operational domains, including preventing risky changes from reaching production, detecting unauthorized configurations, accelerating incident response times, optimizing rollback decisions, and identifying configuration drift.
As enterprises continue to grapple with increasingly complex technical stacks and the growing integration of autonomous agents into daily operations, the fundamental question of what changed and whether it triggered an incident remains difficult to answer. Empirik is positioning itself as the definitive system of record to answer that question for human operators today, and for the autonomous agents that will increasingly manage those systems tomorrow.

Comments (0)