Increasingly, artificial intelligence is playing key roles in driving business value, boosting efficiency while reducing costs. This trend is driving up demand for tech talent capable of building and maintaining AI models, which calls for a mix of evolving software coding capabilities and soft skills. We spoke to five leading artificial intelligence experts, to gauge how best to gain entry into a career in the space.
Highest-earning software salaries ranked — A deep dive into the highest-earning salaries that are being offered across the software market.
The AI job market
With so many areas of artificial intelligence being developed today, it makes sense that job creation in the space is bound to continue evolving with the space.
Szymon Idziniak, machine learning engineer at STX Next, suggests splitting the AI job market into a few key roles:
- Machine learning engineer – Someone who has a strong IT/engineering background, and the ability to write whole software by themselves. “They should also have strong machine learning background and need to understand how something is working before starting to use it.”
- ML/deep learning researcher – Someone who has a strong mathematics and machine learning background. “They should have the ability to develop new neural networks from scratch to solve difficult programs. They will often lack some knowledge of software engineering, but could be supervised by another developer,” said Idziniak. Additionally, this role is mainly focused on just creating and evaluating new models.
- Data scientist – Someone who has background in statistics, IT and mathematics. “They often lack software engineering background, but is good at analysing data and creating reports. They will have a good ability to write in notebook and provide useful insights of data, uses more classical ml approaches like XGBoost, scikit-learn, Pandas and NumPy.”
- MLOps Engineer – Someone who has a strong DevOps background with knowledge of how to write code and also understands basic machine learning approaches. “They will mostly focus on how to utilise cloud providers in real projects.”
The best new roles in AI
When it comes to determining the job roles that are set to drive the most value from artificial intelligence-powered capabilities, experts see much of this value-add to come from product development pertaining to customer service and compliance.
“Prompt engineering is a role that has emerged recently with the development of generative AI technology. Prompt engineers are responsible for designing and refining the prompts or inputs that are used to generate text or other outputs from AI models,” says Kunal Purohit, chief digital services officer at Tech Mahindra.
“Another role that I can see emerging is AI product managers, who will be responsible for developing and managing generative AI products, from ideation to launch. The AI conversation designer, meanwhile, can be another role who is an expert in designing conversational interfaces for AI-powered chatbots, virtual assistants, and voice-activated systems, ensuring smooth and engaging user interactions.
“As the unprecedented power of AI and generative AI comes with a great deal of security and responsibility issues for its use, an ethics engineer will play a vital role to ensure transparency, fairness, and avoiding any unintended biases in the generated output. Then, the AI trainer will also get into the mainstream career path as they will have responsibility of training generative AI systems, teaching them to recognise patterns and generate output that aligns with specific goals.”
As every industry explores ways to harness generative AI to improve their business operations, many industry specific roles may emerge and evolve with time, and as generative AI continues to unfold its potential.
AI-led business processes – getting the balance right between business impact and staff satisfaction — David Semach, CEO of AIM Reply, spoke to Information Age about how businesses can keep staff motivated and productive alongside AI-led business processes.
Programming languages
According to Martin Butler, professor of management practice at Vlerick Business School, Python will be king when it comes to entering into and maintaining a career in artificial intelligence, due to it being widely used and easy to learn.
He explains: “It provides an excellent entry into programming for novices, yet it is still used in complex scenarios by very seasoned programmers. Experienced developers will learn to add additional languages, often context-specific, but Python is the best entry point that can take you very far on the AI developer journey.”
When considering which programming languages not worth learning, Butler cites typical scripting languages like JavaScript, PHP, Perl and Ruby: “These are all very accomplished scripting languages which are powerful in their own right, but not necessarily for AI applications.
“The exception? Phyton, since it can be used for both scripting (i.e. easy to learn) and programming (more suited to AI work).”
Top technical qualifications
When looking to apply for job roles focused on AI, it’s important not to underestimate the challenges that the exciting but demanding technologies can bring — especially to those new to the space.
“To put this in perspective, the UK’s main AI research organisation, DeepMind, used to recruit the top computer science graduate from Cambridge University each year, but they stopped that practice several years ago,” said Clare Walsh, director of education at the Institute of Analytics.
“They found that even the most promising, hardworking 21 year old was unable to cope with the demanding environment of experimental AI. As a company policy, DeepMind now require a PhD qualification as a minimum to get a job there.”
To become a chief data officer — a position increasingly involving work in AI — it is likely that a degree in a relevant field would be required as a minimum. Walsh continued: “There are different routes to do that. For example, to apply for Chartered data scientist status, a gold standard in the field, you’ll need a degree as a minimum.
“There are, of course, many people currently employed who went through different routes. Realistically, this current ‘AI summer’ or period of rapid expansion of AI only began in earnest around 2016. For the first few years, you had to be a PhD researcher, or perhaps completing an MSc to have access to university training in these fields.
“However, the field is becoming more formally regulated and although there will be older, more experienced coders who have no formal training, it would be risky to assume that the best jobs will still be available to anyone in, say 10 years’ time.”
UK data science degrees and diplomas — Here, we list out nearly 100 different data science university degrees (Bachelors and Masters), plus post-graduate diplomas and certifications.
Retraining as an AI specialist
One last question remains: is it ever too late to retrain as a specialist in artificial intelligence?
“I don’t think it’s ever too late,” said Heather Dawe, UK head of data at digital consultancy UST. “Diversity in the people who develop and use AI is so important – people who have done other things previously in their career or who have lived and worked in other ways bring diverse experiences and knowledge with them and this is a strength in itself.
“Further to this, given the uses of AI are likely to become more pervasive, we will all come into increasing contact with it. I think society will in turn become more aware of this and this familiarity will hopefully encourage more people to consider retraining as their interest is piqued.”
Related:
Five top tech companies hiring this week — The UK ended 2022 as Europe’s largest and most powerful tech hub, and this is reflected in the job market.
