By As cloud infrastructure becomes more complex, so do the demands on engineering teams. The traditional toolset — dashboards, alerts, and manual intervention — is proving unsustainable. But what if the cloud could manage itself, like a self-driving car?
That’s the core thesis behind Sedai, an autonomous cloud management platform founded by Suresh Mathew, who previously led large-scale cloud efforts at PayPal and eBay.
📹 Going on record for 2026? We're recording the TFiR Prediction Series through mid-February. If you have a bold take on where AI Infrastructure, Cloud Native, or Enterprise IT is heading—we want to hear it. [Reserve your slot
Beyond automation: what autonomy really means
“Automation plus decision making is autonomous,” says Mathew. Where most tools rely on human-triggered workflows, Sedai makes its own decisions — when to scale, restart, optimize, or repair services — based on real-time metrics and safety policies.
Mathew draws a key analogy: “The moment you put a driver in the driver’s seat, it’s not a self-driving car anymore. The same goes for cloud. If a person has to click a button, it’s not autonomous.”
Autonomous action leads to measurable results. For one of the world’s largest cybersecurity firms, Sedai executed over 100,000 production operations with zero incidents, saving $3.5 million in just six months.
The system improves availability by detecting early warning signals in latency, traffic, and resource usage — and taking action before issues escalate. “It’s not trying to save money — it’s removing waste as a first-class priority,” explains Mathew.
Freeing engineers to build
Sedai isn’t about replacing people. It’s about changing the work they do. In large organizations, optimization often becomes a tug-of-war between FinOps and engineering teams. Sedai removes 80% of that low-level toil.
“Engineers are supposed to be builders, not cleanup crews,” says Mathew. “We let them focus on what matters.”
Sedai supports SaaS and on-prem deployments, with wide compatibility across Kubernetes, containers, VMs, serverless, and storage. Critically, it doesn’t require access to logs or PII. The system runs on metrics alone, with strict controls and knobs for security teams to manage behavior.
It also integrates LLMs for developer queries — not decision-making. Users can ask Sedai’s interface natural-language questions like, “What app is consuming the most cost?” or “Which deployment caused this spike?”
Looking forward
While many competitors claim to be autonomous, Mathew says Sedai is one of the few that actually operates without human triggers — and takes full responsibility for safety. With nine patents and a clean track record, the company is pushing toward a future where cloud environments tune themselves.
