Cloud Engineer Career Roadmap

You execute well-defined work. Someone tells you what to build and you build it, asking good questions along the way. The expectation is competence and reliability, not independent judgement.

• Provision and configure resources from a specification
• Work confidently in the console and on the command line
• Follow runbooks and escalate appropriately
• Write basic automation and read other people's

You own components. Given a problem rather than a specification, you produce a solution and take responsibility for it running. This is where most of the market sits and where the work becomes genuinely interesting.

• Design and implement a service or subsystem end to end
• Define infrastructure as code and own the pipeline that deploys it
• Participate meaningfully in on-call and incident response
• Understand cost implications of your choices and act on them
• Review other engineers' work usefully

You own systems and influence how the team works. The technical work gets easier; the ambiguity gets harder. Much of your value is now in deciding what not to build.

• Design systems spanning several services and teams
• Make and defend trade-offs under real constraints
• Lead incident response and run post-mortems that produce change
• Mentor others deliberately rather than incidentally
• Write design documents others can act on

The path forks. Staff and principal engineering roles stay close to implementation while operating across the organisation. Architecture roles focus on design, standards and direction, often with less day-to-day coding.

• Set technical direction across multiple teams
• Identify problems nobody has articulated yet
• Balance technical quality against delivery pressure credibly
• Communicate with non-technical stakeholders effectively
• Build consensus rather than issuing decisions

Neither branch is more senior than the other. Choose based on whether you want to stay close to the code or move towards design and direction — plenty of people move back and forth.

• Own something end to end. The single fastest way to grow. Volunteer for the thing nobody wants to own.
• Be on call. Nothing teaches system behaviour like being responsible for it at three in the morning.
• Write things down. Design documents, post-mortems, runbooks. Writing is how influence scales beyond conversation.
• Learn one adjacent discipline properly. Networking, security, data or reliability — depth in one adjacent area is worth more than shallow breadth across all of them.
• Change environments occasionally. Five years of varied experience beats one year repeated five times.