ICT CAREER GUIDE CLOUD ENGINEERING
What Does a
Cloud Engineer Do?
Skills · Jobs · Career Path
Discover the real work behind cloud systems, follow three engineering cases, and build your first practical cloud project.
BUILD · OPERATE · SECURE
Reliable services start with deliberate engineering.Quick answer
What does a cloud engineer do?
A cloud engineer is an IT professional who builds, configures, secures, monitors and maintains cloud computing infrastructure. They help applications run reliably, investigate failures, manage access, automate changes and control costs. The job combines networking, operating systems, security, scripting and practical problem-solving.
- Build infrastructure
- Protect systems
- Resolve incidents
- Automate work
What does a cloud engineer actually work on?
Imagine a school where students check examination results through an online portal. A developer writes the program that identifies a student and retrieves the correct result. But the portal also needs somewhere to run, a protected database, working connections, monitoring and a recovery plan. A cloud engineer helps provide and maintain that operating environment.
A simplified request follows this route:
Student’s phone → Internet → Cloud-hosted application → Protected database → Result displayed
The engineer does not necessarily build the application’s features. Their responsibility usually centres on the infrastructure and operational systems that support them. In a smaller organisation, a developer and cloud engineer may be the same person.
Five building blocks you will encounter
| Building block | What it does | School portal example |
|---|---|---|
| Compute | Provides the processing power to run applications | Processes a student’s results request |
| Storage | Holds files, documents and backup objects | Keeps authorised downloads and backups |
| Networking | Connects services and controls traffic | Links the application to approved services |
| Database | Organises structured records | Stores student registrations and results |
| Identity and access management | Controls what people and services can do | Limits which accounts may view or change records |
On a small screen, scroll sideways within the table to view every column.
Managed cloud services can supply some of these functions without the customer maintaining physical servers. If you are new to this subject, read NIT’s Cloud Computing Explained for Beginners for the underlying concepts; this article focuses on the engineer’s work and career .
The cloud engineer focuses on the environment that lets an application run. A developer focuses on its features. Both roles may overlap, especially in smaller teams.
Seven responsibilities of a cloud engineer
1. Prepare infrastructure
Before an organisation launches a service, the engineer helps choose and configure suitable compute, storage, networking and access controls. A small public website might need managed static hosting. A financial application may require carefully separated environments, protected data services, monitoring and more extensive recovery planning.
The job is not to buy the largest server available. It is to match the infrastructure to the application’s needs.
2. Deploy and update applications
Deployment means placing software in its intended operating environment so authorised users can access it. Cloud engineers may support automated deployment pipelines, manage environment settings, verify a release and coordinate recovery if a change causes problems.
For example, a release might accidentally connect an application to the wrong database environment. The team needs to detect the mistake, restore the correct configuration through an approved procedure and test the whole user workflow.
3. Monitor performance and availability
Monitoring tells engineers what is happening before they guess at the cause of a problem. Useful signals include availability (can people reach the service?), response time (how long does a request take?), error rate (how often do requests fail?), resource utilisation (how much compute and memory are used?) and logs (what events were recorded?).
4. Troubleshoot incidents
When a portal fails, an engineer checks what is affected, when it started, what changed recently and what the evidence shows. The investigation normally follows confirm → observe → narrow down → correct → verify → document . Restarting a server or increasing its capacity without diagnosing the problem may waste time or create new risks.
5. Keep systems secure
Cloud engineers help implement least-privilege permissions, multifactor authentication, restricted network connections, secure credentials, encryption, logging and tested recovery arrangements.
A cloud provider secures certain parts of its services, while customers remain responsible for particular settings and data protections. The exact division depends on the service used. AWS calls this the shared responsibility model . [1]
6. Automate repeated work
Manually recreating the same environment ten times can introduce inconsistent settings. Infrastructure as Code lets engineers describe resources in files that can be reviewed and reused. Terraform is one tool for planning and managing such changes. Automation still requires careful review. [2]
7. Manage costs, reliability and recovery
Cloud services may charge according to usage. Engineers review resource utilisation, remove unnecessary services and assess whether capacity is appropriate. They also help design backups and recovery procedures that are tested rather than assumed to work.
Build
Prepare the computing, network and storage resources an application needs.
Protect
Control access, secure credentials and review risky configurations.
Operate
Monitor services, investigate errors and verify recovery.
Improve
Automate repeated work, refine performance and manage unnecessary costs.
A typical day in cloud engineering
Not every cloud engineer spends the day doing the same tasks. Some work on launching new systems; others spend most of their time improving and supporting services already in use.
| Part of the day | Illustrative activity | Why it matters |
|---|---|---|
| Morning | Review overnight alerts, errors and open incidents | Establish what needs attention |
| Mid-morning | Investigate a slow application | Find the cause before changing resources |
| Late morning | Discuss a deployment with developers | Confirm technical and business requirements |
| Afternoon | Review a network or automation change | Prevent unsafe configurations |
| Later afternoon | Examine operating costs and system health | Identify waste and emerging problems |
| Handover | Record evidence and unresolved issues | Keep the next person informed |
On a small screen, scroll sideways within the table to view every column.
Cloud engineers also explain incidents, document configurations, discuss risks and sometimes participate in on-call arrangements. Their work is as much about technical judgement and communication as it is about command-line tools.
Scenario 1: A school’s admission portal fails
The problem: Students can open the admissions webpage but receive errors when submitting applications. The school needs the service restored without exposing private records or losing submissions.
Follow the engineer’s investigation
- Confirm the scope. Test an authorised sample submission. The homepage loads, but the form fails. This narrows the problem to part of the application, not necessarily the entire website.
- Collect evidence. Review monitoring and logs. Suppose the application reports a database authentication failure after an approved credential rotation.
- Test the hypothesis. Check whether the application is using the correct secret reference and whether the database accepts its authorised connection. Do not paste passwords into public logs.
- Apply an approved correction. Update the affected configuration or secret reference through the organisation’s change procedure. Restart or redeploy only if required.
- Verify the complete workflow. Submit a fictional test application, confirm the expected database record and check that the confirmation appears. Also check that earlier records remain intact and error rates return to normal.
- Prevent recurrence. Document the incident and add a post-deployment connection check to the credential-change procedure.
Evidence to record: failing form, relevant redacted log, confirmed root cause, authorised change and successful end-to-end test.
The lesson is that removing an error message is not the same as proving the service recovered .
Example diagnostic log (fictional)
A redacted log might show that the homepage still responds while the form-submission service cannot authenticate to its database:
Engineers verify outcomes: successful submission, correct database record, normal error levels and preserved previous records—not merely an absent error message.
Scenario 2: An online shop becomes slow
The problem: A shop launches a promotion. Visitors increase, pages take longer to load and customers abandon purchases. The owner suggests purchasing a larger server.
Before adding capacity, the engineer compares several measurements.
| Observation | What it might suggest | Next investigation |
|---|---|---|
| High CPU usage | A demanding application process or insufficient capacity | Inspect processes and request patterns |
| Slow database queries | Indexing or query-performance problem | Examine query metrics and execution plans |
| Failed requests increase | A service or dependency may be failing | Review application and infrastructure logs |
| Network timeouts | Connectivity or configuration problem | Check network paths and service health |
| Unexpected cost increase | Unused resources or unusually high usage | Review the resource inventory and billing |
On a small screen, scroll sideways within the table to view every column.
Suppose the evidence identifies a poorly performing database query. The engineer works with developers to test a correction, repeats an equivalent workload test and checks whether the response time and failure rate have improved.
Decision rule: Measure first, diagnose second, make a controlled change third, and verify afterwards. More servers are not always the answer.
Scenario 3: A SACCO tests its backups
The problem: A fictional SACCO wants to know whether financial records can be restored after accidental deletion or a serious service failure. The engineer must establish realistic recovery requirements and demonstrate that backups work.
Two terms guide the work:
- Recovery Point Objective (RPO): the maximum acceptable period of data loss, measured in time.
- Recovery Time Objective (RTO): the target time to restore a service after disruption. [3]
Work through a recovery test
Suppose the SACCO agrees on a 30-minute RPO and a 90-minute RTO . A simulated incident occurs at 4:00 pm . The newest recoverable backup contains data up to 3:15 pm , and the service is restored at 5:10 pm .
| Measurement | Calculation | Result | Target met? |
|---|---|---|---|
| Data recovery gap | 4:00 pm − 3:15 pm | 45 minutes | No: exceeds 30-minute RPO |
| Service restoration | 5:10 pm − 4:00 pm | 70 minutes | Yes: within 90-minute RTO |
On a small screen, scroll sideways within the table to view every column.
The engineer needs to improve the backup or recovery design because the recovery point did not meet the target, even though the restoration time did. The team must also check restored-data consistency, encrypted access, retention and reproducibility.
These figures are illustrative recovery targets , not universal recommendations. A real institution must agree on targets that fit its needs and test that they can actually be met.
How to choose what to investigate first
A skilled engineer does not start with the most dramatic possible explanation. They begin with evidence that narrows down the failure.
| Symptom | Useful first checks | Avoid assuming |
|---|---|---|
| The entire website is unreachable | Service status, DNS, connectivity, recent changes | A restart is always required |
| The homepage works, but forms fail | Application logs and database connections | The whole network is broken |
| The site slows during busy periods | Traffic, response time, resource measurements | More servers are the only answer |
On a small screen, scroll sideways within the table to view every column.
Use a simple incident note for each investigation: problem → impact → evidence → cause → approved correction → verification → prevention .
This routine helps learners explain decisions to other people and gives employers useful evidence of disciplined troubleshooting.
Essential skills and tools
Cloud engineering combines infrastructure knowledge with hands-on problem solving. You do not have to learn every advanced platform feature at once.
| Skill | Why it matters | Beginner demonstration |
|---|---|---|
| Linux and operating systems | Understand processes, permissions and services | Inspect a log and running process |
| Networking | Understand IP addresses, DNS, ports and HTTPS | Explain how a browser reaches a service |
| One cloud platform | Learn compute, storage, networking and access | Deploy a supervised learning resource |
| Security | Protect accounts, networks and data | Explain least-privilege access |
| Scripting | Automate small routine tasks | Write a Python or shell script |
| Git | Record changes to source and configuration | Keep a repository with descriptive commits |
| Monitoring | Detect errors and performance issues | Interpret sample metrics and logs |
| Troubleshooting | Solve problems from evidence | Diagnose a controlled failure |
| Documentation | Make work reproducible and reviewable | Write a clear incident or project report |
On a small screen, scroll sideways within the table to view every column.
Microsoft’s Azure Administrator guidance identifies familiarity with operating systems, networking, servers, identity, compute, storage and monitoring as important infrastructure knowledge. [4]
Do cloud engineers need programming?
Basic scripting is valuable for automating repeated tasks and processing information. Python, Bash and PowerShell are common choices. Some positions require substantial coding; others focus more heavily on administration and infrastructure operations. The important thing is understanding the commands and scripts you run.
What should you learn later?
After the fundamentals, explore Docker (application containers), Terraform (Infrastructure as Code), GitHub Actions (deployment automation), Kubernetes (container orchestration) and cloud-native monitoring tools. Learning the tools without understanding networking and systems makes troubleshooting much harder.
Beginner project: Deploy a cloud website with Azure
You will publish a harmless demonstration webpage using GitHub and Azure Static Web Apps , change it, observe an automated deployment and investigate a practice failure. The steps follow Microsoft’s official portal quickstart. [5]
This is a managed-hosting introduction , not a full production cloud-infrastructure qualification. It teaches a useful first workflow without creating a publicly reachable virtual machine.
Before starting
You need a GitHub account, an Azure account with an eligible subscription, and permission to use the account. Use only fictional data. Never publish credentials, actual student records or other confidential information. Azure Static Web Apps offers a Free plan, but other services may incur charges. [6]
Follow the deployment workflow
-
Create a repository.
Open the
official starter template
and create a repository called
my-first-static-web-app. The starter includes asrcfolder. A public repository is suitable only for non-sensitive demonstration files. - Open Azure. Sign in at portal.azure.com , search for Static Web Apps and choose Create .
-
Set the resource details.
Choose your eligible subscription, create a dedicated resource group named
cloud-learning-lab, select the Free plan, and use GitHub as the deployment source. Select themy-first-static-web-apprepository and itsmainbranch. -
Configure the starter template.
Choose
Custom
for the build preset. Use
./srcas App location , leave API location empty, and set Output location to./src. These values match the named template, not every possible application. - Deploy. Review the configuration and create the resource. In GitHub, open the repository’s Actions tab. Confirm that the deployment workflow reports success.
- Verify the live page. Open the URL shown in Azure. Confirm that the webpage loads over HTTPS and displays the expected sample content. Workflow success and website availability are two separate checks.
-
Change the page.
Edit
src/index.html, change the heading to My First Cloud Deployment , and commit tomain. Watch the next workflow run, then refresh the website and verify the new heading. - Investigate a practice failure. If your new content does not appear, inspect the workflow status, repository branch, commit history and build paths. For a controlled demonstration, use a disposable test copy to set an incorrect path, read the error, restore it and verify recovery. Never deliberately break a real business system.
- Review security and spending. Ensure the repository contains no secrets and Azure contains only the resources you intended. Where available, set a budget alert. A budget alert is not a spending cap ; it does not automatically prevent usage charges. [7]
- Clean up. Delete the Static Web App when finished. Delete the dedicated resource group only if it contains nothing else you need. Review resource and billing information afterward.
Troubleshooting worksheet
| Check | Expected result | If not, investigate |
|---|---|---|
| GitHub commit | The change appears in commit history | Whether the edit was committed |
| Workflow trigger | A new Actions run appears | Repository and branch configuration |
| Deployment | Workflow succeeds | The failing step and logs |
| Build paths |
The paths match the
src
structure
|
App and output folder settings |
| Website | HTTPS page displays updated content | Correct URL and published version |
On a small screen, scroll sideways within the table to view every column.
GitHub’s documentation shows how to inspect individual workflow runs and their logs. [8]
What to put in your project portfolio
Write a short report that another learner can follow. Include the goal , the architecture (GitHub → deployment workflow → Azure Static Web Apps → browser), screenshots of successful deployment, the change test, a redacted troubleshooting example, the security/cost precautions and the cleanup result.
If you cannot explain the components and reproduce the outcome, repeat the exercise before describing it as a completed skill.
Learning outcome: you have practised a repository-driven deployment, update verification, controlled troubleshooting and resource cleanup.
Build a portfolio that demonstrates engineering judgement
After your first deployment, complete more challenging projects in safe learning environments.
| Portfolio project | Skills demonstrated | Evidence an employer can inspect |
|---|---|---|
| Managed cloud deployment | Hosting, version control and release verification | Working demonstration and deployment record |
| Monitoring and incident drill | Logs, metrics and troubleshooting | Synthetic failure, investigation and verified fix |
| Backup recovery or Infrastructure as Code | Recovery testing or reproducible configuration | Test results, reviewed files and clear documentation |
On a small screen, scroll sideways within the table to view every column.
Keep proprietary information, real passwords, private network details and unauthorised data out of public repositories. Each project should describe what you built, why you chose the design, how you tested it and what you would improve .
How to become a cloud engineer: Two routes
There is no single educational path. Some engineers start with operating systems and networking; others begin as software developers.
Route A — through IT and infrastructure
Computer fundamentals → Linux and networking → IT support → Cloud operations/administration → Cloud engineering
This route is useful for someone who enjoys systems, service monitoring and troubleshooting.
Route B — through software development
Programming → Applications and databases → Deployment → Cloud automation and operations → Cloud engineering
This route is useful for someone who enjoys building software and wants to take responsibility for running it reliably.
Seven progression checkpoints
| Checkpoint | Learn | Proof of readiness |
|---|---|---|
| 1 | Computers and networking | Explain an end-to-end web request |
| 2 | Linux and scripting | Read logs and automate a small task |
| 3 | One cloud platform | Explain compute, storage, networking and permissions |
| 4 | Deployment | Publish and test a demonstration app |
| 5 | Monitoring and security | Investigate a controlled incident safely |
| 6 | Automation or recovery | Show repeatable changes or a tested restore |
| 7 | Supervised experience | Document technical tasks and collaborate with a team |
On a small screen, scroll sideways within the table to view every column.
The goal is demonstrable competence, not completing a fixed number of months. A person with networking experience may progress differently from someone beginning with basic computer skills.
Cloud engineer jobs in Kenya: Employer expectations
Kenya’s Standard Classification of Occupations (KeSCO-2022) identifies Cloud Engineer as occupation 2543-16 under Cloud Computing Professionals. This recognises the occupation; it does not measure the number of available jobs or prove a hiring growth rate. [9]
Real job specifications provide more useful insight into employer requirements.
Jubilee Life Insurance — Cloud Engineer
A 2026 job specification includes cloud infrastructure, monitoring, automation, security, cost management and recovery. It required at least three years of relevant hands-on experience and a related university degree; certifications were desirable. The deadline was 25 February 2026, so it is a historical example , not a currently open position. [10]
APA Life Assurance — Cloud Engineer
A Nairobi listing dated 5 November 2026 covers infrastructure, operations, security, governance, performance and cost management. It is described as a middle-level role, rather than a position automatically suitable for a first-time learner. [11]
Identigate — DevOps & Cloud Engineer
An employer description includes AWS, Linux, networking, CI/CD pipelines, Infrastructure as Code and monitoring. Its published requirements include a relevant degree and at least one year of experience in a similar role. [12]
What should beginners search for?
Look at cloud support , infrastructure support , IT operations , junior systems administration and suitable deployment-support roles as well as cloud engineering vacancies. The title alone does not establish that a job is entry-level. Read its academic and experience requirements carefully.
What about salary in Kenya?
Compensation depends on responsibilities, employer, experience, specialisation and working arrangements. Compare recent Kenyan postings that disclose pay; check whether the amount is monthly or annual, gross or net, and whether allowances or on-call work are included. A converted US salary is not a reliable Kenyan salary estimate.
Cloud engineer vs DevOps, developer and architect
| Role | Main emphasis | Example responsibility |
|---|---|---|
| Cloud engineer | Infrastructure and operational reliability | Configure cloud networking and monitor services |
| DevOps engineer | Software delivery and automation | Build reliable deployment pipelines |
| Backend developer | Application logic and APIs | Create the service handling admissions |
| Cloud architect | Overall cloud-system design | Plan the infrastructure and its trade-offs |
| Cloud security engineer | Cloud-specific protection and risk | Review permissions and sensitive configurations |
On a small screen, scroll sideways within the table to view every column.
The boundaries overlap. A cloud engineer may automate deployments, and a DevOps engineer may provision cloud infrastructure. Read job descriptions instead of relying entirely on titles. See NIT’s What Does a Backend Developer Do? for the application-development perspective.
Cloud certifications: Which should you consider?
A certificate can provide a structured learning target, but it does not replace hands-on skills or meet every employer’s experience requirements.
| Platform | Certification | Suitable preparation |
|---|---|---|
| AWS | Certified Cloud Practitioner | Foundational AWS/cloud concepts |
| AWS | Certified CloudOps Engineer – Associate | Cloud operations; AWS describes a candidate with roughly one year of related experience [13] |
| Microsoft Azure | Azure Administrator Associate (AZ-104) | Intermediate administration of identities, networks, compute, storage and monitoring [4] |
| Google Cloud | Associate Cloud Engineer | Google lists no prerequisite but recommends 6+ months of hands-on platform experience [14] |
On a small screen, scroll sideways within the table to view every column.
AWS Certified CloudOps Engineer – Associate succeeded the previous SysOps Administrator – Associate certification. [13]
Choose one platform first , complete projects, and then check the provider’s current exam outline, fees, eligibility and renewal conditions before registering.
Is cloud engineering the right career for you?
Cloud engineering may suit you if you enjoy investigating technical faults, understanding how systems communicate and keeping online services dependable. It also requires careful judgement: a poorly planned change can interrupt a service, expose information or increase costs.
If you enjoy building visible website features, software development may be a more natural first direction. If you enjoy defending systems and examining security risks, cybersecurity may appeal more. The skills overlap, so your first field need not determine your entire career.
AI-assisted tools can help explain logs and draft configuration changes, but those suggestions still need checks for accuracy, permissions, security and cost. A professional must understand and verify what gets deployed.
Frequently asked questions
Can I become a cloud engineer without a degree?
You can learn the skills without a university degree, and some employers consider equivalent experience. Other employers specify a degree or particular qualifications. Always check the individual job advertisement.
Can I start after KCSE?
Yes. Begin with computers, networking, operating systems, programming basics and introductory cloud services. Practical projects and further training can help you progress towards supervised technical roles.
Is cloud engineering difficult?
It combines several technical subjects, but learning step by step makes it more manageable. Start with networks and systems before advanced automation and architecture.
Does cloud engineering need advanced mathematics?
Basic numerical reasoning helps with capacity, performance and cost calculations. Some specialist fields require deeper mathematics, but most beginner cloud administration exercises do not.
Can cloud engineers work remotely?
Some roles can be performed remotely through approved management systems. Others require office attendance, controlled working environments or on-call cover.
How long does it take to become a cloud engineer?
There is no universal timeline. Your current skills, available practice, training and targeted job requirements matter more than a promised number of months.
Do I have to learn AWS, Azure and Google Cloud together?
No. Master the foundations on one platform before learning another. Common ideas such as networking, identity, security and monitoring transfer across providers.
Will AI replace cloud engineers?
AI can assist with scripts, documentation and troubleshooting suggestions, but technical accountability, incident investigation, security review and verifying changes remain critical.
Where should a beginner start today?
Focus on three achievable actions: understand how applications and networks work , complete a small cloud deployment and verify it , and build deeper projects involving monitoring, security, automation and recovery .
A strong cloud engineer is not defined by how many platform features they can name. The skill is understanding the requirement, choosing a suitable solution, making a controlled change, and showing clear evidence that the system works.
For learners choosing an ICT starting point, NIT’s published Full-Stack Software Development course covers deployment topics and industrial attachment as part of broader software training. [15] See the official admissions information to apply for a suitable programme. [16]
Build your ICT foundation
Start Practical ICT Training at Newton Institute of Technology
Interested in cloud engineering? NIT’s Full-Stack Software Development programme teaches programming, databases, Git, testing and software deployment. Its curriculum introduces Docker and AWS or Google Cloud deployment as part of broader software training. The programme consists of eight months of institutional learning followed by a compulsory attachment of at least three months.
These skills can provide a foundation for further specialised cloud engineering learning. Explore the published programme and apply through NIT’s official admissions page.
Sources and references
Primary technical, government, employer and institutional sources. Employment examples are dated records, not claims that the positions remain open.
- [1] Amazon Web Services — Shared Responsibility Model . Cloud provider and customer security responsibilities.
- [2] HashiCorp Developer — Terraform Core Workflow . Infrastructure as Code planning and change management.
- [3] AWS Well-Architected Framework — Disaster Recovery Objectives . Definitions of recovery point and recovery time objectives.
- [4] Microsoft Learn — Azure Administrator Associate . Official Azure administration knowledge and certification details.
- [5] Microsoft Learn — Create Your First Azure Static Web App . Official portal quickstart and GitHub deployment instructions.
- [6] Microsoft Learn — Azure Static Web Apps Hosting Plans . Free and Standard hosting plan details.
- [7] Microsoft Learn — Create and Manage Azure Budgets . Budget-alert behaviour and limitations.
- [8] GitHub Docs — Using Workflow Run Logs . Inspecting workflow results and failure logs.
- [9] Government of Kenya — Kenya Standard Classification of Occupations (KeSCO-2022) . Cloud Computing Professionals and Cloud Engineer occupation classification.
- [10] Jubilee Life Insurance — Cloud Engineer Job Specification, 2026 . Kenyan employer duties, qualifications and experience.
- [11] Corporate Staffing Services — APA Life Assurance Cloud Engineer Vacancy . Nairobi 5 November 2026 middle-level recruitment example.
- [12] Identigate — DevOps & Cloud Engineer . Employer-published infrastructure and automation requirements.
- [13] AWS — Certified CloudOps Engineer, Associate . Official AWS operations certification and renamed title.
- [14] Google Cloud — Associate Cloud Engineer Certification . Certification content and recommended experience.
- [15] Newton Institute of Technology — Full-Stack Software Development Course . Published course pathway, deployment topics and industrial attachment.
- [16] Newton Institute of Technology — Admissions . Official admissions and registration information.
Technical and employment sources reviewed 10 November 2026. The three workplace scenarios use fictional organisations and illustrative data.