By Laily UPN 
In the high-stakes world of modern cloud computing, system downtime is more than a technical nuisance—it is a catastrophic financial and operational risk. For years, the industry standard has been reactive: wait for a system to break, trigger an alert, and scramble to identify the root cause. However, a new startup incubated within the halls of Sequoia Capital is looking to flip that script, moving from reactive fire-fighting to proactive, autonomous prevention.
Empirik, an observability startup that tracks system changes and predicts their ripple effects across complex digital environments, officially announced its spin-out from Sequoia on Tuesday. The company revealed it has secured $21 million in seed funding, backed by a consortium including Sequoia, Canapi Ventures, and Alumni Ventures. By functioning as an autonomous “traffic cop” for infrastructure, Empirik aims to redefine how engineers manage the increasingly fragile, hyper-connected architectures of the modern enterprise.
The Genesis: A Problem Born from Experience
The origins of Empirik are rooted in the veteran experience of its founders. Avon Puri, who joined Sequoia Capital in 2020 as Chief Digital and Information Officer, brought with him over a decade of hands-on experience managing infrastructure at industry heavyweights Rubrik and VMware.
Three years ago, as the true potential of large language models (LLMs) began to materialize, Puri and his colleague, Sequoia IT leader Sudheer Dhurjati, identified a glaring gap in the market. They observed that while developers were using AI to write code faster, the infrastructure engineers responsible for keeping that code running were still buried under manual, error-prone, and reactive tasks. They realized that the traditional observability model—which treats system failures as inevitable—was fundamentally flawed. They posited that if an AI could understand the complex web of dependencies within a massive tech stack, it could predict, and therefore prevent, outages before they ever reached the end user.
This insight became the blueprint for Empirik. The goal was to build a tool that does not just monitor logs and metrics, but understands the logic of system changes and their potential cascading impacts.
Chronology: From Concept to Independent Venture
The trajectory of Empirik has been rapid, marked by a deliberate incubation period under the mentorship of one of the world’s most prestigious venture firms.
- 2021–2022 (Conceptualization): Puri and Dhurjati spend months mapping out how AI agents could interpret infrastructure telemetry and system state changes. They focus on the concept of "dependency mapping"—understanding how a single update to a database or a configuration file might crash a downstream service.
- 2023 (Incubation): Sequoia Capital formally incubates the project. During this phase, the team builds the core architecture of the platform, testing it against the high-demand environments of Sequoia’s own portfolio companies.
- Early 2024 (Leadership Transition): To prepare for a market launch, Sequoia recruits Kartik Chandrayana, formerly the Chief Product Officer at Quantum Metric and a Vice President of Observability at Salesforce. His mandate is to scale the product into a standalone business.
- October 2024 (Market Entry & Funding): Empirik formally announces its independence from Sequoia. The company closes its $21 million seed round, signaling strong institutional confidence in its ability to disrupt the $20+ billion observability market.
The "Traffic Cop" Paradigm: How Empirik Works
At the heart of Empirik’s value proposition is its ability to act as an automated gatekeeper. Bogomil Balkansky, a partner at Sequoia Capital, characterizes the current state of infrastructure management as a “money pit,” where companies spend exorbitant amounts on tools that provide data but lack the intelligence to synthesize it.
"Most existing observability tools fail to understand complex system dependencies," Balkansky noted in a recent interview. "They tell you the system is on fire, but they don’t tell you who set the match."
Empirik changes this dynamic through three distinct layers of automation:
- Risk Assessment: Every proposed change to an infrastructure environment is scanned for potential ripple effects.
- Autonomous Guardrails: For routine, low-risk changes, Empirik provides the green light, allowing the system to update autonomously.
- Human-in-the-Loop Intervention: For complex or high-risk updates, the platform flags the change for human review, providing the engineer with a detailed report on why the change might be dangerous, essentially acting as an intelligent assistant that does the "homework" before the human makes a final decision.
Implications for the Future of DevOps
Empirik’s launch arrives at a critical juncture for software engineering. As generative AI and agentic coding tools like Cursor and Claude Code accelerate the velocity of software development, the infrastructure layer is buckling under the weight of unprecedented change.
"What agentic AI did for software, Empirik wants to do for infrastructure engineering," says CEO Kartik Chandrayana.
The implication is a shift in the role of the Site Reliability Engineer (SRE). Historically, SREs have spent a significant portion of their week in "on-call" rotations, troubleshooting incidents that are often the result of minor configuration drift or human error. By offloading these routine tasks to an autonomous agent, Empirik allows SRE teams to transition from manual maintainers to high-level system architects.
This is not about replacing engineers; it is about scaling their capabilities. As software becomes more complex, the cognitive load on engineers is reaching a breaking point. By automating the "traffic control" of the data center, companies can maintain the high pace of innovation required by modern markets without sacrificing the stability of their underlying systems.
Supporting Data and Market Reception
Though Empirik is in its early stages as an independent entity, its proof-of-concept phase has yielded significant results. The platform has already onboarded a diverse array of clients, ranging from high-growth startups to Fortune 500 stalwarts.
Notable clients include Guardant Health, a leader in precision oncology, and a major player in the consumer packaged goods (CPG) sector. These deployments suggest that Empirik’s value proposition is universal—whether in highly regulated fields where downtime is a safety risk, or in high-volume industries where downtime is a direct hit to the bottom line.
In terms of the competitive landscape, the observability market is currently saturated with incumbents like Datadog, New Relic, and Dynatrace. However, Empirik’s leadership argues that their focus on proactive change intelligence rather than reactive metrics creates a defensible "category of one." They position themselves not as a replacement for current monitoring stacks, but as a complementary "intelligence layer" that works alongside existing AI SRE platforms such as Resolve and other Sequoia-backed initiatives like Traversal.
A New Era of Autonomous Operations
The $21 million funding round is a clear signal that investors are betting on the "agentic" future of the enterprise. As we move deeper into the era of AI, the infrastructure that powers our digital economy must become as intelligent as the applications running on top of it.
For Empirik, the path forward is clear: continue to refine the model’s understanding of global system dependencies and expand the platform’s integration capabilities. By effectively creating an "autonomous infrastructure engineer," the team is addressing one of the most stubborn inefficiencies in modern tech.
As software development continues to accelerate, the gap between "code speed" and "infrastructure stability" will only widen. Empirik aims to bridge that divide, ensuring that as the world moves faster, it doesn’t break more often. In the words of its leadership, the goal is simple: to make the invisible work of keeping the internet running not just easier, but smarter.
rifanmuazin 
Suro Senen 