
Picture a tester with four years of experience. They write sharp bug reports, read API responses without trouble and know how releases go wrong. Now they want to move into development, cloud, security or AI, and one question stalls them: which of these four years count, and which do I have to redo?
You can change IT specialisation without starting from zero, but only if you work out what carries over. Some skills move straight across, others need a refresher, and a few need real retraining. Guess wrong and you either relearn what you already know or walk into interviews missing something basic.
This guide is a general framework, not a roadmap for one role. It covers how to choose a target, which skills tend to transfer, a five-stage process, how to build experience before your title changes and how to explain the move to recruiters. For any one switch, it links to the detailed MicroDegree guides.
Most people don't switch on a whim. Knowing which trigger is yours decides what kind of move makes sense.
Going deeper in your current specialisation is also a legitimate plan. The World Economic Forum's Future of Jobs Report 2025 found that employers expect about 39% of workers' core skills to change by 2030, so some learning is coming whether or not you change domain.
Usually, yes, because your experience isn't one lump. It has four parts, and they carry over differently.
| What you have | What it means | Does it carry over? |
|---|---|---|
| Transferable technical skills | Debugging, reading logs, Linux basics, version control, SQL, scripting | Mostly yes, across almost every specialisation |
| Domain-specific knowledge | Test design techniques, ITIL ticket handling, front-end frameworks, database tuning | Partly. The thinking helps, but the details are tied to the old domain |
| Practical work experience | Shipping releases, working with teams, handling incidents, dealing with deadlines | Yes. Employers value it, though it needs translating |
| Newly required competencies | Tools, concepts and responsibilities that the target role has and yours didn't | No. These are the real gap, and the rest of this guide is about sizing and closing it |
A tester understands the software lifecycle, bug triage and often some automation. A support engineer has troubleshooting habits and a feel for how systems fail. A backend developer knows APIs, databases and application design.
Nobody skips the gap. The tester still has to learn production-grade code, the support engineer infrastructure as code, the backend developer model evaluation. What they skip is the beginner stage.
Chasing whatever is trending is the usual way to choose badly. Score two or three candidates against the same six questions instead, and write your own answers next to each row.
| Factor | Question to ask yourself |
|---|---|
| Technical strengths | Which parts of my current job do I do faster than my peers? |
| Work you enjoy | On a good day at work, what was I actually doing? |
| Responsibilities | What would I own in this role: on-call, design, testing, data, security reviews? |
| Learning requirements | How many new concepts sit between me and an entry-level hire in this role? |
| Long-term interest | Would I still want this work in five years if the trend faded? |
| Internal movement | Does my company have this kind of team, and has anyone moved into it? |
| Current specialisation | Possible target | Transferable skills | Additional learning required |
|---|---|---|---|
| Technical support / service desk | Cloud engineering, system administration | Troubleshooting, OS and networking basics, incident handling | Cloud platform fundamentals, scripting, infrastructure as code |
| Manual / software testing | Test automation, SRE, AI testing | Test thinking, bug analysis, SDLC knowledge | Programming, CI/CD, monitoring, or model evaluation, depending on the target |
| Front-end development | Full-stack development, cloud | JavaScript, browser debugging, Git, build tools | Server-side code, databases, deployment and hosting |
| Backend development | Platform engineering, AI application development | APIs, databases, application architecture | Infrastructure and delivery tooling, or LLM integration and evaluation |
| Database administration | Data engineering, cloud database engineering | SQL, backup and recovery, performance tuning | Pipelines, distributed data tools, managed cloud database services |
| DevOps | Platform engineering, SRE | CI/CD, containers, infrastructure as code | Internal developer platforms, reliability targets and incident practice |
| IT operations | Cybersecurity, cloud operations | Systems administration, monitoring, access management | Security tooling and threat concepts, or cloud service operations |
Already a few years in and unsure what to learn first for cloud? Read Cloud Skills to Learn After Your First IT Job before any beginner course.
The same five stages work whether you're heading to security, data or DevOps.
List the tools you've used, the projects you've worked on and what you owned in each. Add what people rely on you for, such as clear incident updates or untangling a vague request. Mark each item as tied to your domain or useful anywhere; the second group is your bridge.
Read ten or fifteen real job descriptions for the role and copy the repeated requirements next to your Stage 1 list. Watch the verbs: "design", "operate" and "investigate" describe very different jobs. The NICE Workforce Framework for cybersecurity does the same, describing roles as groups of tasks plus the knowledge and skills they need.
Then aim at a realistic first role, often an adjacent one such as QA automation on the way to SRE. If you can, ask someone who does the job what their week looks like.
Split the gaps into two piles: skills almost every posting asks for, and tools that appear in only a few. Learn the first pile and ignore the second until a project or an interviewer forces the issue. Nobody needs every tool in the industry.
Keep the plan to two or three skills and one thing you can show at the end.
Recruiters judge what they can see. Evidence can be a GitHub project, a script that now runs on your team's server, a short write-up of an incident you helped fix, or an internal tool. Take on a small responsibility at work that touches the new domain. A certification gives a quick signal, but it works best beside practical proof.
Start inside your own company. CompTIA's guidance recommends exploring an internal transfer first, and internal candidates already know the systems and people. Talk to your manager about where you'd like to head and what a move would need. Internal moves face the same standard as outside hires, so Stage 4 evidence matters here too.
Without an internal route, apply to adjacent roles, rewrite your resume around the target role and prepare a short explanation of why you're switching.
Your title lags behind what you can do. Build experience in the new area while you still hold the old title:
Real exposure settles what no article can: whether you like the work.
A recruiter skimming your resume asks one thing: is this move deliberate? Lead with your strongest overlap, show a project from the new area and present the old experience as an asset. Say "I worked in testing for four years, and the part I enjoyed most was building the automation framework", not "I'm tired of testing."
When asked why you want to change your specialisation, answer in three parts: what you did, what drew you to the new area and what you've done about it. For example:
"I've spent four years in QA, mostly on API and regression testing. Over the last year I kept being pulled into pipeline failures, so I started maintaining our CI jobs and built a small deployment script for our staging environment. I'd like to move into a role where I own that work full time, and my test background helps me catch problems before release."

There's no honest universal answer, so distrust any plan that promises a move in 30, 60 or 90 days. It depends on:
Treat the first month as a test. After an audit, a few job descriptions and one small piece of evidence, you'll have a better estimate than any generic timeline.
Your earlier years still count when you change your IT specialisation. A good transition starts with a clear list of what you know and what you still need, then adds proof that you can do the new work. Still choosing? Shortlist two options and test them with a small project or a week of shadowing.
MicroDegree teaches in Kannada first, with instructor-led live online classes, practical capstone projects and morning and evening batches for working professionals, covering Cloud, DevOps and Generative AI. It's one option among several, and it can't stand in for the experience-building steps above.
Once you've picked a direction, you'll need a place to build the missing skills. You can look through MicroDegree's e-learning courses in one place and choose the ones that match your next career move.
Not at the start. Do your audit and a first small project quietly, then raise it once you can show something. "Here's what I've built and where I'd like to grow" goes better than "I want to leave this team."
Sometimes, especially where your direct experience is thin. It depends on the company and role. Compare the new role's level with the scope you hold now.
Ten years of experience is an asset if you can show current, hands-on work in the new area. Senior people often do best in roles that combine the old and new domains.
It can if each move looks random. Two connected switches over a decade read very differently from four unrelated ones in four years, so keep a clear thread you can explain.
Run a small experiment on both: two weeks on a starter task or a day of shadowing each. Pick the one where progress felt natural, not the one that sounds more impressive.