Career tracks
Five directions, twelve roles. Tap a track to see what those jobs really do all day.
Data track5 roles+
The most beginner friendly on-ramp in tech. It runs from answering questions with data, to building the dashboards, to laying the pipes, to designing the platform, to predicting the future. If you like patterns, business, and explaining things clearly, start here.
Data Analyst
A Data Analyst answers business questions using data. Someone asks why did sales drop in the Western Cape last month, and the analyst pulls the numbers, cleans them, finds the pattern, and explains it in plain language with a chart or a short report. Think of the analyst as a detective who works with a spreadsheet and a database instead of a magnifying glass.
How it differsThe analyst answers questions and reports on what already happened. A BI Developer builds the reusable dashboards and data models the analyst and the whole company then use. A Data Scientist goes further and predicts what will happen next. The analyst mostly looks backward and explains, the scientist looks forward and forecasts.
Core skills
Starter certifications
How to break inThis is the easiest tech job to break into with no degree. Learn SQL and Power BI for free, then build three small projects on public South African data, load shedding stats from Eskom, crime stats, or Takealot style sales you invent, and put them on a GitHub or a simple portfolio site. Pass DP-900, it is affordable and quick to study in two to four weeks. Then apply for junior analyst and data captured roles, look on OfferZen, LinkedIn, and Pnet. The ExploreAI Academy in Cape Town trains exactly this pipeline and lists Data Analyst as a first job title on their programme.
BI Analyst or BI Developer
A Business Intelligence Developer builds the dashboards and the semantic model that sits under them. Where the analyst answers one question, the BI Developer builds a self service tool so the whole company can answer that kind of question themselves. They design the data model, write the calculations once so everyone uses the same definition of revenue, and make dashboards that a busy manager understands in five seconds.
How it differsA Data Analyst uses dashboards and answers ad hoc questions. The BI Developer builds and maintains the dashboards and the model behind them, so it is a step more technical and more permanent. A Data Engineer feeds clean data into the BI layer, the BI Developer shapes that data into a business friendly model. Analyst consumes, BI builds the reporting layer, engineer builds the pipeline underneath.
Core skills
Starter certifications
How to break inStart as a Data Analyst, then go deep on Power BI and DAX, because DAX is where most people stall and it is your differentiator. Rebuild your analyst portfolio projects as proper models with reusable measures, not just one off charts. Many South African firms run on Power BI because it comes bundled with the Microsoft 365 they already pay for, so PL-300 is genuinely in demand. Pass PL-300, then apply for BI Developer and Reporting Analyst roles, which pay noticeably more than pure analyst roles.
Data Engineer
A Data Engineer builds the pipelines that move and shape data. Every night, or every few minutes, data has to travel from apps, tills, and websites into a place where analysts and scientists can use it, cleaned and organised. The engineer builds and runs that plumbing so it is reliable, fast, and does not break at month end. If the analyst is the chef, the engineer stocks the kitchen and keeps the water running.
How it differsAnalysts and BI Developers consume the data the engineer prepares. The Data Engineer rarely makes the charts, they make sure the data is there, clean, and on time. A Data Architect designs the blueprint of the whole platform, the engineer builds and operates the pieces. The line is design versus build and run, though at small companies one person does both.
Core skills
Starter certifications
How to break inThis role usually is not a first job, it is a strong second step. Come in as a Data Analyst or a junior software developer, learn Python and cloud properly, then move across. Build a real pipeline project, pull data from a public API on a schedule, clean it, load it into a cloud warehouse, and show the code on GitHub. Pass DP-700, which is the certification path Brendon himself walked. Data Engineer is well paid, PayScale South Africa put the average near R460 000 per year in 2026, so the effort pays back.
Data Architect
A Data Architect designs the whole data platform. Before anyone builds a pipeline or a dashboard, the architect decides where data lives, how it flows, which tools connect to which, how it stays secure and POPIA compliant, and how it will still work when the company is ten times bigger. They draw the master plan and set the standards everyone else follows. This is the town planner of the data city.
How it differsA Data Engineer builds and runs pipelines inside the plan. The architect draws the plan and picks the tools and the rules. Where an engineer thinks about this pipeline, the architect thinks about every pipeline for the next five years, plus cost, security, and governance. It is the most senior and most strategic seat in the Data track.
Core skills
Starter certifications
How to break inThere is no beginner path here, and be wary of anyone selling you one. You reach architect after several years as a strong Data Engineer or Solutions Architect, once you have felt the pain of systems that were designed badly and learned what good looks like. For a beginner the honest advice is aim here as a five to eight year goal, and start by becoming an excellent Data Engineer first.
Data Scientist
A Data Scientist builds predictive and machine learning models. Instead of only explaining the past, they use maths and code to forecast the future or find hidden patterns, which customers are likely to leave, is this transaction fraud, what should we recommend next. They run experiments, test hypotheses, and turn messy data into a model that makes a decision or a prediction.
How it differsA Data Analyst explains what happened, the Data Scientist predicts what will happen and often quantifies the uncertainty. A Machine Learning Engineer takes the scientist's model and turns it into a reliable production system that serves millions of requests. So the scientist leans toward maths, statistics, and experimentation, the ML engineer leans toward software engineering and deployment. Many jobs blur these two.
Core skills
Starter certifications
How to break inThe maths bar is real, so this is harder to enter cold than analysis. Two honest routes for a South African beginner. First, a structured programme like ExploreAI Academy in Cape Town, which teaches Python, statistics, and machine learning in a project based way and lists Data Scientist as an outcome, and which offers bursaries up to R70 000 for citizens. Second, self study the maths and Python, then win a couple of Kaggle competitions and publish the notebooks. Either way, expect twelve to eighteen months of serious study. Many people wisely start as an analyst, earn while they learn the maths, then cross over.
Cloud and Infrastructure track2 roles+
The people who run the digital land the software lives on. If you like reliability, automation, and the satisfaction of a system that just works and scales, this track pays very well and is in heavy demand, because every company is moving to the cloud.
Cloud Architect or Solutions Architect
A Cloud or Solutions Architect designs how an application will run in the cloud. Given a business need, a booking app that must handle a rush without falling over, they decide which cloud services to use, how the pieces connect, how it stays secure and affordable, and how it survives a data centre problem. They produce the technical blueprint that the developers and DevOps engineers then build.
How it differsA DevOps Engineer automates and operates the systems day to day, the architect designs them up front. A Data Architect focuses specifically on data platforms, the Solutions Architect covers the whole application, compute, networking, storage, and security. Architect is a design and decision role, DevOps is a build and run role, and at small firms one person wears both hats.
Core skills
Starter certifications
How to break inYou do not start as an architect, you grow into it. But the cloud track has a friendly first rung, the fundamentals cert. A beginner can pass AZ-900 or AWS Cloud Practitioner in a month, which opens junior cloud support and cloud administrator roles. From there, get the AWS Solutions Architect Associate, gain two or three years of hands on experience, and the architect title follows. Cloud skills add a real premium, OfferZen 2026 noted that AWS, Azure, or Kubernetes certifications often lift pay by ten to twenty percent.
DevOps Engineer
A DevOps Engineer automates the path from a developer's laptop to live users, and keeps the whole thing running. They build the pipelines that test and ship code automatically, manage the servers as code so nothing is set up by hand, monitor for problems, and get the team back up fast when something breaks. The name joins Development and Operations, because the job is to make those two work as one smooth machine.
How it differsA Software Developer writes the application, the DevOps Engineer builds the machinery that ships and runs it reliably. A Cloud Architect designs the system, the DevOps Engineer automates and operates it. Think designer versus mechanic and pit crew. In South Africa, where load shedding and connectivity make reliability a daily fight, good DevOps people are prized.
Core skills
Starter certifications
How to break inTwo common doors. Either come from software development and drift toward automation and infrastructure, or come from IT support and system administration and learn to code your infrastructure. For a beginner, learn Linux and Git deeply, automate something real, a small app that deploys itself when you push code, and show it on GitHub. Get a fundamentals cert then the Terraform Associate. DevOps sits among the best paid entry points, OfferZen 2026 found entry level DevOps engineers earning meaningfully more than backend developers.
Software track1 role+
The builders. If you like making things that people use, apps, websites, and features, and you enjoy the loop of write it, run it, fix it, this is the broadest and most flexible track in tech, and the most forgiving of a non traditional background.
Software Developer or Software Engineer
A Software Developer writes the code that becomes the apps, websites, and systems people use every day. They turn an idea, let users book a doctor online, into working software, then test it, fix the bugs, and improve it. The two titles overlap heavily. In practice, developer often points at writing features, while engineer suggests a broader concern with how the whole system is designed, tested, and scaled, though many companies use the words interchangeably.
How it differsA DevOps Engineer ships and runs what the developer writes. A Data Engineer moves data rather than building user facing features. A Machine Learning Engineer is a software engineer who specialises in getting models into production. The plain Software Developer builds the product the customer actually clicks on, which makes it the most visible and often the first coding job people land.
Core skills
Starter certifications
How to break inThis is the classic no degree success story. Pick one path, front end web is the gentlest start, learn it properly, and build real projects that solve real problems, a site for a local spaza shop, a booking tool, a small SaaS. Your GitHub and a live portfolio are your degree here. In South Africa, WeThinkCode teaches software development for free and hires without requiring matric maths in the usual way, and OfferZen is built to match self taught developers with employers. Entry level developer pay in Cape Town averaged about R23 846 a month in early 2026 per OfferZen, and remote work for overseas companies can multiply that.
Security track1 role+
The protectors. As every business moves online and POPIA makes data protection the law, the people who defend systems are in short supply and high demand. If you like puzzles, thinking like an attacker, and being the calm one in a crisis, this is your track.
Cybersecurity Analyst or Security Engineer
A Cybersecurity Analyst watches over an organisation's systems and defends them from attack. They monitor for suspicious activity, investigate alerts, patch weaknesses before criminals find them, and respond when something goes wrong, a phishing attack, a ransomware attempt, a leaked password. A Security Engineer goes further and builds the defences themselves, secure networks, firewalls, and identity systems, rather than only watching them.
How it differsThe analyst mostly monitors, investigates, and responds, the closest thing to a digital security guard and detective. The engineer designs and builds the protective systems, more like the architect of the fortress. Both differ from a DevOps Engineer, who keeps systems running, while security people keep them safe. In small South African firms one person often covers both analyst and engineer duties.
Core skills
Starter certifications
How to break inMost people do not start in security, they arrive after a year or two in IT support or networking, because you must understand how systems work before you can defend them. A realistic South African path, get CompTIA A+ then Network+ to land an IT support job, learn on the job, then add SC-900 and Security+ and move into a junior security or SOC analyst role. Practise on free platforms like TryHackMe and Hack The Box and show your progress. Note that Microsoft AZ-500 retires on 30 September 2026 and is replaced by SC-500, which adds AI system security, so aim at the current exam.
AI track3 roles+
The newest and fastest moving neighbourhood, and where Brendon himself lives. It grows straight out of the Data and Software tracks. If you are excited by machine learning and by the wave of large language models, this is where data science, software engineering, and product thinking meet.
Machine Learning Engineer
A Machine Learning Engineer takes models and makes them work in the real world at scale. Where a Data Scientist might prove that a model can predict fraud in a notebook, the ML Engineer turns that into a reliable service that scores millions of live transactions, fast, cheaply, and without falling over. They are software engineers who specialise in the messy business of getting models into production and keeping them healthy.
How it differsA Data Scientist explores data and builds the model, the ML Engineer productionises and scales it, so the scientist leans maths and the engineer leans software and systems. A Data Engineer moves data in general, the ML Engineer builds the specific pipelines that feed and serve models. An AI Engineer, in today's usage, often works more with ready made large language models than with training models from scratch. The lines are genuinely blurry and titles vary by company.
Core skills
Starter certifications
How to break inThis is a second or third role, not a first job. You arrive from software engineering or from data science, then add the missing half. A software developer learns the machine learning maths, or a data scientist learns to write production grade code and deploy. For a beginner the honest sequence is, become a solid Software Developer or Data Scientist first, then specialise. It is one of the best paid roles in tech, which reflects how much ground you must cover to get there.
AI Engineer
An AI Engineer builds products and features on top of existing AI models, especially large language models like Claude and GPT. Rather than training a model from zero, they wire powerful ready made models into real applications, a chatbot that knows your company's policies, a tool that reads contracts and flags risks, an assistant that drafts emails. They combine software engineering with techniques like retrieval augmented generation, prompting, and agent design. This is exactly the work Brendon does as a senior AI engineer.
How it differsA Machine Learning Engineer often trains and optimises models, the AI Engineer more often composes and orchestrates existing ones into products, though the roles overlap and are converging. A Software Developer builds general apps, the AI Engineer builds apps whose core intelligence comes from an AI model, which brings new problems, prompts, context, cost per call, and unpredictable outputs. A Prompt Engineer focuses narrowly on crafting the instructions, the AI Engineer builds the whole system around them.
Core skills
Starter certifications
How to break inThe fastest growing door in tech, and unusually open, because almost nobody has ten years of experience in a field this young. If you can already code, you can start today, build a small but genuinely useful AI app, a study assistant, a tool for a local business, and ship it. Brendon's own leap, data analyst to data engineer to senior AI engineer in under six months, shows how fast this track can move for someone who builds relentlessly. Learn one LLM API well, master retrieval augmented generation, and put working projects in front of employers. Proof of building beats any certificate here.
Prompt Engineer
A Prompt Engineer designs the instructions that get the best, safest, most reliable results out of an AI model. They craft and test the wording, structure, and examples that steer a model, measure which versions work better, and build reusable prompt templates and guardrails for a product. In practice this is rarely a standalone job title anymore, it has become a core skill folded into the AI Engineer, content, and product roles.
How it differsAn AI Engineer builds the whole application, including the code, data, and infrastructure, and uses prompting as one tool among many. A pure Prompt Engineer focuses only on the language that goes into the model. Because prompting alone is a thin foundation for a career, treat it as a high value skill to master rather than a destination. It pairs beautifully with the Software, AI, and even content and marketing tracks.
Core skills
Starter certifications
How to break inThe most accessible entry point into AI for a non coder, and the best advice is do not stop here. Use free access to models like Claude to practise daily, document your before and after results, and solve a real problem for a real person or business. Then add either light coding to grow toward AI Engineer, or domain depth to become the AI savvy expert in law, marketing, or education. On its own, prompt engineering is a strong skill and a weak career, so use it as a springboard into the wider AI track.
How to choose between them+
How to choose your track
Before you learn a single tool, get clear on one thing. What kind of work makes you lose track of time. The tech industry is not one job, it is a city of neighbourhoods, and you will do far better living in the one that suits your temperament than chasing the one with the loudest salary.
Here is a simple way to find your street.
- You love asking why and telling the story. You enjoy spreadsheets, patterns, and explaining what the numbers mean to a person who is not technical. Start in the Data track, most likely as a Data Analyst.
- You love making things look clear and clickable. You want a manager to open a dashboard and instantly understand the business. That is BI, still in the Data track.
- You love plumbing and reliability. You care that data or software arrives on time, every time, even during load shedding. Look at Data Engineering or DevOps.
- You love designing the whole system on a whiteboard. You think in boxes and arrows and trade offs. That is the Architect path, in either the Data or the Cloud track.
- You love building the thing people actually touch. Apps, websites, features. That is the Software track.
- You love the maths of prediction and experiments. You want to forecast churn or detect fraud. That is Data Science, and it opens the door to the AI track.
- You love protecting things and thinking like an attacker. That is the Security track.
A few honest truths for a South African beginner. You do not need a degree to start, you need proof that you can do the work, which means projects on GitHub and one or two starter certifications. Salaries in this guide come from the OfferZen 2026 Developer Salary and Benefits Report and PayScale South Africa, verified July 2026, and they move fast, so always check the live source before you quote a number to anyone. Entry level developer pay in Cape Town averaged around R23 846 per month in early 2026 per OfferZen, and it climbs steeply with skill and with remote work for overseas companies. Pick the track you enjoy, because the one you enjoy is the one you will practise long enough to get paid well for.
Sources, verified July 2026: OfferZen Developer Salary and Benefits Report 2026, OfferZen software developer salary, PayScale South Africa Data Engineer, ExploreAI Academy, Microsoft Learn certifications, AWS certification.