Cloud-native systems were supposed to make everything faster and more secure. Instead, many organizations find themselves managing increasingly complex environments where security responsibilities are scattered across teams, tools, and platforms that don't communicate well with each other.
Yugandhar Suthari, has been watching this evolution firsthand. As a senior security engineer at Cisco with twenty years of experience building large-scale systems, he's seen the same patterns repeat across industries—from his six years supporting HBO's streaming infrastructure to his recent work with Sports Illustrated's digital platform expansion.
"We've gotten really good at detecting problems," Suthari says. "But we're still not great at preventing them or understanding why they happen in the first place."
Most security teams today operate platforms they didn't build, using tools they didn't configure, with policies they inherited from previous teams. The result is a patchwork of automated systems that can spot anomalies but struggle to explain what normal operations should actually look like.
Suthari experienced this challenge during his extensive work in media streaming. His platform contributions at HBO supported major streaming experiences, including the infrastructure behind Game of Thrones streaming that Vulture highlighted as a defining moment for digital entertainment. Later, The Wall Street Journal spotlighted scalable infrastructure work behind streaming expansions during his time with Sports Illustrated's launch of Sports Illustrated TV in a context where platform fragility could never remain invisible.
"When you're streaming to millions of users simultaneously, you can't afford to learn about security issues from incident reports," he explains. "The system has to know what it's supposed to be doing and flag when it starts drifting from that intent."
This media industry experience shaped Suthari's approach to what he calls "self-aware infrastructure." Rather than relying on static security rules, his recent research focuses on systems that learn their own normal operating patterns and can identify when something changes.
His work in this area recently led to a presentation on "AI for Threat Detection in Single Page Applications: A Proactive Security Approach" at IIT Indore's 16th International Conference on Contemporary Computing and Informatics, where he demonstrated how machine learning can anticipate security issues in web applications before they become exploitable vulnerabilities.
His adaptive automation framework works by establishing baselines for infrastructure behavior—tracking patterns like API call frequencies, how often secrets get rotated, and changes in permission structures. When the system detects drift from these established patterns, it can trigger corrective actions before problems escalate.
"It's not about generating more alerts," Suthari says. "Most teams are already drowning in alerts. It's about helping the system recognize when it's not behaving the way it was designed to."
This approach has practical applications in areas like MLOps pipelines, where Suthari's open-source tools help teams maintain visibility into both their deployment processes and secrets management—two areas where automation often creates blind spots.
As a senior IEEE judge and reviewer for Engineering Applications of Artificial Intelligence, an esteemed scientific journal, Suthari brings both academic rigor and operational experience to security research. He also serves as a session chair at BIDA 2025, where he sees a persistent gap between security theory and operational reality.
Many organizations can pass compliance audits but struggle with basic operational challenges like rolling back problematic deployments or understanding the scope of a configuration change.
"Real security maturity isn't about having the right certifications," he says. "It's about whether your system can explain what went wrong and help you fix it quickly."
He points to a common scenario: a misconfigured resource that passes initial security scans but creates problems during a rollback attempt. In these situations, checklists and compliance frameworks don't help—teams need systems that can trace the relationships between components and explain the impact of changes.
Suthari's perspective represents a shift in how organizations think about cloud security. Instead of focusing primarily on threat detection and response, he advocates for treating security as a fundamental design consideration.
"If your architecture doesn't reflect what you actually intended to build, you're not really defending your infrastructure," he explains. "You're defending your assumptions about what that infrastructure should be doing."
This means security teams need to work more closely with platform architects and developers, ensuring that security considerations are built into system design rather than layered on top afterward.
As cloud environments become more complex, Suthari believes the organizations that succeed will be those that invest in clarity and accountability from the ground up. This means building systems that can explain their own behavior, making security responsibilities explicit rather than assumed, and ensuring that automation enhances rather than obscures operational understanding.
Cloud security’s future belongs to the architects who assume less, and verify more. In distributed systems, reliability and security converge: both are signals of sound design, and both fail silently when misunderstood.
Suthari’s design lens focuses less on threat modeling and more on architectural intent. “If your system doesn’t reflect what you meant to build, you’re not defending infrastructure. You’re defending assumptions,” he says.
"The future belongs to teams that can build systems they actually understand," he says. "Because in distributed architectures, if you can't explain how something works, you can't secure it effectively."
Security won’t be solved by stacking tools. It will be solved by systems that know what they are, and tell you when they are not. Because in cloud-native design, the system diagram is the threat model. And the real breach isn’t just what you missed, it’s what your architecture refused to show you.
For organizations still struggling with cloud security complexity, Suthari's advice is straightforward: start by making sure you understand what your systems are actually doing. Everything else builds from there.
Because in tomorrow’s distributed architectures, clarity is security, and discipline is the only architecture that scales.