Empirik launched Tuesday with $21 million in seed funding from Sequoia Capital, Canapi, and Alumni Ventures, spinning out of Sequoia after a two-year incubation to build what it calls an autonomous infrastructure engineer. The startup's pitch: predict outages before they happen by tracking every system change across a large environment and modeling how those changes ripple through dependencies. It is targeting the same DevOps and site reliability teams that today spend most of their time reacting to failures rather than preventing them.
The company was founded by Avon Puri, who joined Sequoia in 2020 as chief digital and information officer after more than a decade running infrastructure at Rubrik and VMware, and Sudheer Dhurjati, another Sequoia IT leader. The two began sketching the product roughly three years ago, when large language models started to look capable enough to reason about complex production systems. Sequoia incubated Empirik in 2023 and recruited Kartik Chandrayana — former chief product officer at Quantum Metric and observability VP at Salesforce — as CEO earlier this year.
Empirik functions as what the team describes as an autonomous traffic cop for infrastructure changes. Low-risk updates get waved through, mid-risk changes trigger guardrails, and the most dangerous updates get flagged for human review before they ship. The goal is to offload routine troubleshooting from engineers who are perpetually outnumbered by the systems they maintain.
“There has always been a lot of money spent in keeping systems up and running”— Bogomil Balkansky, Sequoia partner
Key facts
- 01Empirik raised $21M in seed funding from Sequoia Capital, Canapi, and Alumni Ventures.
- 02Sequoia incubated the company in 2023 before spinning it out as an independent business this week.
- 03Early customers include S&P Global, Guardant Health, and a major CPG company.
- 04Founders Avon Puri and Sudheer Dhurjati previously ran infrastructure at Rubrik and VMware.
- 05Kartik Chandrayana, former Quantum Metric CPO and Salesforce observability VP, was recruited as CEO earlier this year.
Sequoia partner Bogomil Balkansky argued that most existing observability tools fail to understand complex system dependencies, and that Empirik is one of the first dedicated tools focused on tracking changes across massive environments. That framing positions the product less as a rival to logs-and-metrics incumbents and more as a new layer that sits above them, reasoning about cause and effect rather than surfacing symptoms after the fact.
Since going live earlier this year, Empirik has signed customers ranging from startups to several Fortune 500 companies, including S&P Global, Guardant Health, and a major consumer packaged goods company. The early traction gives the company a bench of production environments to train against — a meaningful data moat in a category where the value of the model scales with the diversity of the systems it has seen fail.
Chandrayana frames the opportunity through the same lens investors have used for AI coding tools. As AI accelerates the pace of software development, the volume of infrastructure changes accelerates alongside it, and traditional monitoring struggles to keep up. Empirik is betting that the same automation curve that produced Cursor and Claude Code for developers will produce a parallel category for the engineers keeping those developers' code running in production.
“What agentic AI did for software, Empirik wants to do for infrastructure engineering”— Kartik Chandrayana, Empirik CEO
Competition in AI-native site reliability is heating up. Balkansky argues Empirik occupies its own category for now, complementary to AI SRE platforms like Resolve and Sequoia-backed Traversal, which focus more on incident response than on change-driven prediction. Whether that distinction holds as the market matures is an open question — SRE platforms tend to expand into adjacent workflows quickly, and change tracking is a natural next step for any tool that already has read access to production.
There are also open questions about how well change-impact prediction generalizes across radically different stacks. A model that reliably flags risky changes at S&P Global's environment may need substantial retraining to do the same at a fast-moving startup, and false positives in this category are expensive — engineers who get paged for non-events stop trusting the tool. Empirik has not disclosed accuracy metrics or how it measures a prevented outage, which will be the numbers that matter once the marketing settles.
The bet worth watching is whether autonomous infrastructure engineering becomes a standalone category or gets absorbed into the incumbent observability stacks. Sequoia clearly thinks the former, given both its incubation of Empirik and its earlier check into Traversal. If the thesis lands, the next 24 months should produce a cluster of AI SRE tools carving up the workflow the way agentic coding tools carved up the IDE — and the seed round announced this week will look like an early entry point into what could become one of enterprise AI's more durable line items.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
