By Madhanraj J

With the disclosure from Anthropic that AI is generating nearly 100 percent of their internal codebase, the simmering debate on the future role of engineers boiled over, making global headlines. The emergence of Agentic AI-powered vibe coding is driving many to say that the multiple development roles that exist today should be collapsed into a single one, since enterprises will now be able to vibe the entire software stack.

While the prospect is very much on the horizon, it is too soon to talk about doing away with key development stakeholders. This is because while vibe coding is being used extensively at the pilot, proof of concept, and prototyping stages, the technology is further evolving to scale up for production, addressing all of the needs of security or functionality. A recent study found that while nearly 60 percent of organizations are experimenting with AI agents, more than 40 percent believe they are too risky to be adopted at scale.

Traditional system development primarily involves two major roles – cloud engineers who build the underlying platform along with a foundation of base services, including those related to observability, backup, lifecycle management, security, and compliance, and application developers who leverage this platform to create the actual application. The onus of ensuring system performance, security, and compliance is on the platform organization or cloud engineering team, with the application developer mainly responsible for building business functionality. Now, if the application developer uses vibing to build an application, the big question is who will be accountable for security, compliance, and other performance parameters?

A possible solution is to also enable cloud engineers with similar artificial intelligence tools to enable them to continue working alongside application developers, and therefore continue with the responsibility matrix for some time into the future.  Let’s explore this thought.

Needed, a blueprint for consistent deployment

By harnessing the native capabilities of generative AI and agentic AI, organizations can curate the global knowledge and best practices of cloud and technology firms, and create the required automation; however, there could be a risk of not meeting organizational standards and compliance needs, and ending up with undifferentiated technology and architecture. This can be addressed by enabling cloud engineers with AI copilots and intelligent automation so they can create deployable architecture – an infinitely reusable “target state architecture” blueprint with the required performance, security, and resilience capabilities – that application developers can consume to vibe secure and compliant applications at scale. Another benefit of such a blueprint is that it does away with ad hoc responses to developers’ service requests, and ensures consistent deployment, while slashing operational and compliance costs. Last but not the least, it offers an elegant way to maintain the separation of responsibilities, allowing the cloud engineer and application developer to work for the security-compliance-performance promise and functionality promise, respectively.

Preparing for new roles

Agentic AI may not make cloud engineers redundant, but it will change their role. As low-level, repetitive tasks get automated, cloud engineers can further take up more sophisticated responsibilities such as orchestrating intelligent systems, managing complex compliance and governance requirements, and of course, designing new architectures. They will be expected to make critical decisions and exercise judgement, for example, pick between trade-offs, interpret regulatory mandates, and translate business objectives into architectural specifications. Newer roles – such as cloud + AI specialist, cloud governance analyst, etc. – will further start coming up.

While predicting that AI agents would fully automate more development tasks over the next few years, a leading analyst firm said that the creation of AI-empowered software would spark a need for even more skilled engineers. In the near to medium term, cloud engineers will remain very relevant to software development. Although vibing is here, it’s not quite “there” yet to meet all enterprise needs around security, compliance, and performance. The cloud engineer and application developer should work side by side – both enabled with agentic AI tools – with a boundary of responsibility delineating their respective roles. AI will only help elevate the cloud engineer.

 

 

(The author is Madhanraj J, AVP, Cobalt Cloud Solution Strategist, Infosys, and the views expressed in this article are his own)