Imagine you’ve been working as a backend developer for six years. You’re earning a healthy salary and working with a good team. However, every other post you see is saying ‘software engineer is dead’ and you only have one career path to pursue – AI engineer. You’ve got some free time on your hands and you’re wondering if you should retrain or if you’re just panicking.
This is definitely a valid concern. To be honest, neither is as exciting as they make out. There’s a massive amount of overlap between these two roles. Whether you choose to go down the AI engineer route or stay as a software engineer depends on what you want to do and how much of your previous experience will transfer.
Software Engineer
A software engineer will design, develop, test and maintain software. According to the National Careers Service, as a software engineer you will discuss requirements, plan your technical solution, write and test code, document your changes and provide support for the product. The key thing here is that you’ll produce deterministic outputs. This means that for the same inputs you will get the same outputs for your system.
AI Engineer
As mentioned above there is a lot of overlap between software and AI engineers. The difference is that with AI engineers, your system will contain a model and models are not deterministic. Most employers in the UK will use the term AI engineer to describe someone who will take an existing model (often a large language model) and turn it into something that can handle traffic, data and failures in production. This can involve building retrieval pipelines, evaluation systems, monitoring costs and lots of data work. Training a model from scratch is a much more niche role and will usually have a different title.
Essentially, an AI engineer is a software engineer who has to work with uncertainty. If you can deliver reliable software then you’ve got most of the hard work done.
Skills
Both require you to have a good knowledge of version control, testing, system design, and being able to explain trade-offs to non-technical people. You’ll also need to know python for AI engineering. Whilst a software engineer might spend their whole career using Java, C# or Typescript.
With AI engineering you’ll need to know some additional skills. These include practical LLM skills, vector databases, fine tuning and a deep learning framework like PyTorch. The biggest difference between the two roles is evaluation. Whilst you can get away with saying ‘it looks like it works’ when choosing a button colour, you can’t do this with a system answering customer questions. Knowing how to evaluate whether your outputs are good is key to knowing whether you can deliver or just demo.
Data
The most useful data we have on how AI affects roles is from PwC’s AI Jobs Barometer 2026. Overall, roles that are ‘professionalises’ by AI (those that require more human expertise) are growing approximately twice as fast as those that are made easier for non-experts. They’re also seeing 42% faster wage growth since 2021. Both software and AI engineers are on the right side of this.
However, there are some concerns. The skills required for the most AI exposed roles are changing over twice as quickly as those with the least exposure. This has increased by 75% in just one year. Entry level roles in these exposed sectors are also seven times more likely to require traditionally senior skills. Overall, the real question should be whether you’re learning fast enough to keep up with the changes to your job description. This applies to both a software engineer who isn’t using any AI tools and an AI engineer who isn’t keeping up with their framework.
Salaries
From government guidance, software developers earn around £30k at entry level and up to around £75k with experience. AI engineers earn around £35k at entry level and up to around £75k. This is a national average and will cover a wide range of seniorities.
At a higher level, you’ll often see AI roles advertised at £55k to £80k mid level and £85k to £130k senior. In London you’ll often see salaries 15-25% higher than the national average. If you’re a specialist within a well funded scale-up you can earn a lot more than this. This is just a snapshot. Salaries are moving very quickly in this market so check live roles within your region and seniority level.
Overall trends
Software Engineer with additional AI skills
This is the best value option for most people. You’ll keep your seniority and gain the AI skills whilst on the job. You’ll become the go-to person in your team for delivering features with models. Many of the best AI engineers have come from this route.
Lateral move to AI Engineer
If you’re drawn to the evaluation, data and experimentation side of AI engineering then this could be for you. There might be a dip in confidence and potentially title as you’ll be junior in the AI specific areas. More and more employers are looking for software engineers with real portfolios rather than just a certificate. If you want a real structured route then the Level 6 Machine Learning Engineer apprenticeship is a valid option.
Specialist
If you want to go down a specialist route, like MLOps, applied research, security or domain specific (healthcare/finance) there will be less people available to replace you meaning salaries will increase.
Tech Recruitment Agency
Whilst a tech recruitment agency won’t make decisions for you, a good agency will get you the quickest insight into what the market wants. Specialist agencies will know about roles weeks ahead of them going live. They can tell you whether AI engineer at a certain company will be a research role or a fancy API integration. They can also tell you what the budget is.
Ask them what skills are coming up in their briefs. Be cautious of any tech recruitment agency that can’t tell you how many AI/software engineers they placed last year and where. Usually this means they’re a CV blasting agency and don’t have any real specialist knowledge.
If you’re hiring, an agency that challenges you to break up a bloated AI engineer role into two clearer roles are helping you.
Getting started
If you’re a software engineer and are interested in AI then don’t quit anything. Try and spend a weekend building something with real data and force yourself to create an evaluation. If you enjoy this then you know what you need to do. If you’re bored then you’ve saved yourself from retraining.
If you’re earlier in your career then make sure to get the engineering fundamentals first. Most people will have gotten to these roles by learning the AI tools as they go.
Talk to someone who can see real hiring data. Talking to a specialist tech recruitment agency is free and will give you more insight into your next realistic step.
Common Questions
Are AI engineers paid more than software engineers?
At the highest level yes, often. However at entry level there’s not a huge difference. Official data shows similar salaries for both roles. The increase comes from the specialist nature of the role and how scarce they are.
Will AI replace software engineers?
So far there’s evidence to suggest that roles will change. There is more judgement, evaluation and system design required. Those who use the tools well will be in a better position.
Do I need a masters to be an AI engineer?
No. Most roles will care more about your demonstrable engineering skills than your post graduate qualification. However, research heavy employers may vary.
How long does it take to move from software to AI engineer?
If you have a good foundation as a software engineer then 6-12 months of focused work to build a portfolio.
Should I use a tech recruitment agency?
Yes! Apply directly if you really admire the company but using a tech recruitment agency will give you insight into the market. They will have roles that are never advertised.
