Startups say no dedicated DevOps

X user Dominus Kelvin argued on May 23 that young startups do not need dedicated DevOps specialists and should instead rely on a senior or mid-level full-stack engineer to run infrastructure work. In the post, Kelvin said that engineer should cover deployment, CI/CD, servers, monitoring and databases rather than adding a separate DevOps hire. The post, published under X ID 2058465029926916155, circulated over the last 48 hours as software engineers debated startup staffing and the scope of “full-stack” work. The thread reflects a familiar argument in early-stage companies: whether infrastructure specialization is premature when headcount and cash are constrained. ### What exactly did the post say? Post 2058465029926916155 framed the issue in direct hiring terms. Kelvin wrote that “young startups” do not need dedicated DevOps specialists and said a “cracked senior or mid-level full-stack engineer” can handle the core operational work instead. The list in the post named deployment, CI/CD, servers, monitoring and databases as responsibilities that should stay with that engineer. (x.com) ### Why does this argument keep resurfacing at startups? Early-stage startups often compress roles because small teams need to ship product before they build functional silos. Kelvin’s post fits that pattern by treating operations work as part of general engineering coverage rather than a standalone specialty. The claim is not that those tasks disappear; it is that one experienced engineer can absorb them during the company’s first phase. (x.com) The staffing logic behind that view is straightforward. A startup with only a handful of engineers may prioritize product development and customer-facing features over a separate platform or site-reliability function, especially if cloud tooling reduces some of the setup burden that once required a dedicated systems administrator. That interpretation is consistent with the wording of Kelvin’s post, which focused on who should do the work, not on whether the work matters. (x.com) ### What work is being grouped under “DevOps” here? Kelvin’s list bundled together several categories that startups often split only later. Deployment and CI/CD refer to how code is built, tested and shipped. Servers and databases cover the infrastructure that runs the application and stores its data. Monitoring refers to the systems that track uptime, errors and performance once software is live. (x.com) By naming those functions together, the post used “DevOps” less as a formal job title than as a package of operational responsibilities. That is one reason the thread drew attention: engineers often agree on the tasks but disagree on when those tasks justify a dedicated specialist. ### Who is the post aimed at? The wording “young startups” makes the target audience narrow. (x.com) Kelvin did not describe larger companies with multiple products, strict compliance requirements or round-the-clock reliability teams. The post instead addressed companies at an early stage, where one experienced engineer may still be expected to move across application code, deployment pipelines and production support. That distinction matters because the claim is about timing. The post argues against making DevOps a separate early hire, not against infrastructure work itself. In practice, the debate usually turns on scale: traffic, complexity, regulatory obligations and how costly downtime would be for the business. ### What happens next in this debate? The next step is visible on X itself. (x.com) Post 2058465029926916155 remains the focal point for replies, reposts and counterarguments from engineers discussing when a startup should split operations from product development. As of May 24, the thread is still live under that post ID, and any follow-up from Kelvin or other participants is likely to appear there first.