Guest: Rob Hirschfeld (LinkedIn)
Company: RackN
Show Name: An Eye on AI
Topic: AI Infrastructure

AIOps is no longer just about monitoring systems and running models — it’s becoming a core discipline of infrastructure management. In this clip, Rob Hirschfeld, CEO and Co-Founder of RackN, explores where AIOps is heading and how it’s beginning to merge with the world of platform engineering and infrastructure automation.

Today, when most people talk about AIOps, they mean managing AI clusters — keeping systems online, feeding jobs, and maintaining models. Hirschfeld believes that’s only half the picture. The next wave, he says, is AI infrastructure operations: building, patching, and automating the physical and virtual environments that AI workloads depend on.

That shift represents a new maturity in AI adoption. While hyperscalers have long automated their GPU clusters and data fabrics, most enterprises are just starting that journey. “We see a wave coming,” Hirschfeld explains, “where enterprise users will have to think about how they operationalize AI infrastructure.” RackN is already working with customers to bridge that gap, helping them adopt automation practices originally developed for hyperscale systems.

This new form of AIOps is as much about mindset as technology. It requires thinking of infrastructure automation as a living system — one that must adapt to new hardware, frameworks, and workloads at unprecedented speed. Instead of treating AI infrastructure as static, organizations need continuous processes for patching, resetting, and scaling — all without downtime.

Hirschfeld also points toward the next frontier: AI managing infrastructure itself. RackN is experimenting with ways to use LLMs and AI coding assistants to help operators script, configure, and automate tasks. Early experiments, he notes, are promising — particularly in what he calls “vibe coding” or AI-assisted operations, where human operators guide AI systems to automate workflows faster.

But Hirschfeld draws an important distinction. Building automation with AI — letting LLMs create or maintain infrastructure logic — is much harder than using AI to assist humans in managing infrastructure. The latter, he says, is where the real progress is happening today. By embedding AI as an abstraction layer, operators can gain new visibility, generate code faster, and troubleshoot complex systems more effectively.

As the boundaries between AI, automation, and operations blur, Hirschfeld believes we’re heading toward a future where infrastructure management is partly deterministic and partly stochastic — combining precise automation with AI-driven adaptability. For RackN, that’s an opportunity to redefine the intersection of human and machine collaboration in infrastructure operations.

Ultimately, AIOps won’t replace platform engineering — it will evolve alongside it. The goal, Hirschfeld concludes, isn’t just smarter automation, but an ecosystem where AI helps engineers manage complexity at every layer of the stack.