DevOps Engineer

Also known as: DevOps Specialist, Cloud Engineer, Infrastructure Engineer, Build and Release Engineer, Cloud Infrastructure Engineer.

What does a DevOps Engineer do?

A DevOps Engineer is an engineer responsible for the path code takes from commit to production and for the infrastructure that code runs on. Their job is to let developers ship changes often and without fear while production stays stable and does not cost more than it should. Typical tasks:

• build and maintain CI/CD pipelines;
• describe infrastructure as code and manage cloud, containers and Kubernetes;
• set up monitoring, logging and alerts;
• automate routine operations that used to be done by hand;
• watch infrastructure cost and find where it can be reduced.

When does a company need a DevOps Engineer?

A company needs a DevOps Engineer when it has more than five to ten developers and shipping is no longer a one-evening job: releases break, environments drift apart, nobody knows exactly how production is configured. Common triggers: a move to the cloud or to Kubernetes, the first large customer with availability requirements, a cloud bill growing for no obvious reason. In a startup of three engineers on managed services a dedicated DevOps is usually unnecessary: the infrastructure is run by a Founding Engineer or a strong Backend Developer.

DevOps Engineer vs System Administrator: what's the difference?

A DevOps Engineer differs from a System Administrator in solving infrastructure problems with code and automation rather than by configuring servers by hand. A System Administrator maintains existing machines and services; a DevOps Engineer builds a system in which servers are created and updated automatically and developers ship code without their involvement. What sets them apart from an SRE is focus: an SRE owns reliability in SLO terms, a DevOps Engineer owns the speed and ease of delivery.

Why is it hard to hire a DevOps Engineer?

DevOps Engineers are hard to hire because the title has become very broad and covers people with very different levels of engineering.

• DevOps is the title for both former system administrators who configure servers by hand and engineers who build infrastructure as a product.
• Experience with one cloud or one toolset is often presented as overall level, and the difference only shows in practice.
• Many candidates have maintained someone else's infrastructure but never built one from scratch or run a migration.
• On-call duties not mentioned up front put strong candidates off even after the offer.

What to decide before hiring a DevOps Engineer?

• What hurts most: release speed, stability or infrastructure cost? The main pain defines the profile, and trying to find someone for everything stretches the search.
• What is the infrastructure now: cloud, own servers, Kubernetes? Experience with your own hardware and experience with managed cloud are different candidate pools.
• Build from scratch or develop what exists? The first needs architectural experience, the second needs patience and the ability to work with legacy.
• Is there on-call, and in what format? Better to say so on the first call than to lose a candidate in the final round.

Where to find a DevOps Engineer?

DevOps Engineers most often come from high-load product companies, cloud providers, fintech and e-commerce, and from Backend Developers who moved into infrastructure. A strong profile signal is a described migration or platform with a clear result (how release frequency, recovery time or cost changed) or contributions to infrastructure open source. NEWHR finds such candidates by the results of their work, not by applications to a job post.

How to assess a DevOps Engineer in an interview?

DevOps Engineers are assessed with a review of infrastructure the candidate built and a practical task on the company's stack: a pipeline, a deployment and monitoring. The review covers architecture, incidents and what exactly they automated. On the company side, a lead infrastructure engineer or Team Lead takes part, along with a developer who will use their work.

Strong signals:

• describes how their work changed release frequency or recovery time after failures;
• thinks about how convenient the pipeline will be for developers;
• knows what the infrastructure cost and what they did about that number.

Red flags:

• describes the work as a list of configured tools, with no result;
• proposes Kubernetes for any task without asking about scale;
• makes changes in production by hand and sees no problem with it.

How long does it take to hire a DevOps Engineer?

We present the first candidates for the DevOps Engineer role within 5–7 days of the brief. A team of ten SREs took us about six months of net search time.

Related roles

• SRE — owns reliability in SLO terms, not delivery speed.
• Platform Engineer — builds the internal developer platform at large scale.
• DevSecOps Engineer — builds security checks into the same pipelines.
• MLOps / LLMOps Engineer — builds delivery and infrastructure for ML models.
• SDET — automates testing inside the delivery pipeline.