Loading page…
The Sequoia-incubated startup wants to calculate a proposed change’s reach before it runs, then block or escalate risky work. Its central test is whether enterprises will trust those assessments in production.
Listen to this story
Empirik is positioning infrastructure-change analysis as a pre-deployment control layer: its system builds an operational graph from cloud, identity, CI/CD and SaaS data, models a proposed change’s blast radius, and routes actions by risk. The Sequoia-incubated startup was developed in 2023 before spinning out, and now has $21 million from Sequoia, S32, Canapi Ventures and Alumni Ventures.
Kartik Chandrayana leads Empirik after product and observability roles at Salesforce and Quantum Metric.
The platform can permit low-risk updates, block unsafe ones, and escalate dangerous changes for human approval.
TCBPay reportedly reduced sensitive configuration-change detection from 30 minutes to 10 seconds; the result comes from Sequoia, not independent measurement.
Empirik has emerged from stealth with $21 million to build an AI agent that evaluates infrastructure changes before deployment. The company says the system can calculate a change’s potential impact, block unsafe actions and send dangerous updates for human review.
Avon Puri and Sudheer Dhurjati built Empirik’s initial version inside Sequoia, which incubated the company in 2023. The startup later spun out independently and is now led by Kartik Chandrayana, formerly Salesforce’s VP of Product for Observability and Big Data and chief product officer at Quantum Metric.
The launch disclosure names Sequoia Capital, S32, Canapi Ventures and Alumni Ventures as investors in the $21 million financing.
Empirik says it captures the intent behind a change from a pull request, ticket or pipeline, then projects the proposed mutation across the live environment before deployment. It says governed workflows can assess risk, block unsafe changes and execute approved ones.
Sequoia positions Empirik against conventional observability tools that surface symptoms such as latency, errors and crashed workloads. Its wager is that a map of changes and dependencies can identify the cause earlier and give teams a chance to intervene before an incident.
That is a complementary layer to AI site-reliability engineering platforms such as Resolve and Traversal, Sequoia partner Bogomil Balkansky told TechCrunch, rather than a claimed replacement for them.
Empirik says its software is powering production environments at Guardant Health, Avahi Systems, TCBPay, a Fortune 50 consumer-packaged-goods enterprise and a Fortune 500 financial-data-services leader. Sequoia says TCBPay cut the time to catch sensitive configuration-file changes from 30 minutes to 10 seconds; that is an investor-supplied customer outcome.
The launch materials do not provide an independently measured outage-prevention rate or an error rate for Empirik’s impact calculations. Those measures will determine whether enterprises let the system permit, stop or escalate production work.
Loading discussion...
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.