Machine learning is a specialist aspect of AI (artificial intelligence) concerned with training computers to learn from inputted data without the requirement for further programming for every situation. It involves creating the algorithms (sets of rules) and computer programmes that enable machines to learn in a similar way to humans.
Never a day goes by when we don’t hear something about AI in the news. Despite being associated with certain risks, AI can make all kinds of tasks quicker and more effective, and it has huge potential for the future.
To give you an idea about the scope of AI, here are a few examples of areas in which it is having an impact:
- Writing and other creative work – generative AI tools, like Bard, ChatGPT and Midjourney, are large language models (LLMs) that use deep learning algorithms to identify patterns and connections; they are trained with so much data that they can create new text, images or music in response to specific criteria
- Medicine – AI can, for instance, help diagnosis disease, speed up drug discovery and monitor the spread of infections
- Customer services – think of those chatbots that answer your questions via online chat or over the phone
- Marketing – for example, when you do your online shopping other products you may like are suggested
- Manufacturing – AI is used to automate production, predict when machines need maintenance and detect quality issues
- Security – face-recognition technology, for example, has applications ranging from assisting with police investigations to adding security features to technology
- Farming – AI can help monitor crops, soil and weather conditions, and identify the best time for harvesting
In this article we’ll take a look at careers in machine learning. We’ll help you think about whether you have the right sort of aptitudes, and explore how you can get started.
Jarosław Krawczyk and Igor Zatoń are employed by Digica, an international software solutions company working at the forefront of AI. They kindly explain a little about their careers and give some useful insights.
What would I do if I worked in machine learning?
The first step in any AI project is to be clear about the problem or task that needs addressing and to define clear goals.
As we’re writing about AI, we thought it would be interesting to ask ChatGPT to outline the 5 key tasks involved in machine learning. We advise you to always verify information produced by AI, but this is what it came up with …
- Data Preprocessing:
Cleaning, transforming, and preparing data for analysis is a crucial step. This involves handling missing values, scaling features, and encoding categorical variables.
- Model Selection and Training:
Selecting appropriate machine learning algorithms or deep learning architectures, training them on the data, and optimizing model parameters for best performance.
- Feature Engineering:
Creating new features or selecting relevant ones to improve the predictive power of the model. This involves domain knowledge and creativity.
- Evaluation and Validation:
Assessing model performance using various metrics and techniques like cross-validation to ensure it generalizes well to new, unseen data.
- Deployment and Maintenance:
Taking the trained model and deploying it in real-world applications. Continuously monitoring and updating the model to adapt to changing data and requirements.
These tasks are fundamental to the machine learning workflow and are often iterated upon (ie repeated) to build robust and effective machine learning systems.
Although in this article our focus is on machine learning, people with a range of expertise including data science and software development are required to work on AI projects. Responsibilities and roles often overlap and different employers may use different job titles; both Jarosław and Igor are data scientists, but their roles include machine learning.
“I’m responsible for proposing solutions to problems, primarily based on available data and knowledge acquired from domain experts”, says Jarosław. “I’m currently working on a project focused on classifying objects detected by military radar. With modern data science tools, it’s possible to estimate the type of object being tracked by radar in real-time.”
Igor explains, “I’m responsible for creating solutions for clients using machine learning techniques. I solve a range of issues starting from intelligent analysis of documents using natural language processing to human detection on images using the latest computer vision solutions. I’ve been involved in many projects, mainly for big industrial companies, but the most interesting was an attempt to crack the pirate code using a combination of global optimisation algorithms and natural language classifiers that returned the level of ‘Frenchness’ of the provided text. My job involves much creativity – it’s definitely not boring!”
Where could I work in machine learning?
Specialists in machine learning can work for employers in industries as diverse as pharmaceuticals, logistics, finance and engineering – in fact any public or private sector organisation where AI has (or has the potential to have) an impact. They also work for firms that specialise in undertaking AI projects for clients; many are small, start-up enterprises.
Those with expertise in machine learning also work in consultancy, for research institutes or in government departments and agencies, such as the UK Office for Artificial Intelligence.
TIP: To explore potential employers and opportunities in machine learning, have a look at some of the job vacancies available through recruitment sites and agencies such as Technojobs, Bubble Jobs and Understanding Recruitment.
Would I suit a career in machine learning?
To work in machine learning you need very strong analytical and creative problem-solving skills. As Igor says, “You definitely need to like maths.” The exact technical skills required vary depending on the position, but you need coding ability, knowledge of relevant programming languages (eg Python and Java) and confidence with various software tools.
You must be able to concentrate and pay attention to detail. It’s also important to work well in a team and to be able to communicate clearly – you often have to explain complex information to clients and other stakeholders who are not experts. Some roles call for project and/or people management skills.
It's essential to be prepared to keep up to date with technological advancements through short courses, reading etc. Jarosław explains, “Methods are being developed quickly. Techniques that were used just two years ago might now be outdated.” Igor adds, “Machine learning is a very hot field. Many articles about new algorithms are published daily.”
To see how well matched your strengths, interests and personality are to a career in machine learning you can take the Morrisby Profile; see buy a pass.
How do I get into machine learning?
There’s no one route into machine learning, but employers seek people who are enthusiastic and knowledgeable about AI. Most people who work in machine learning are graduates, although it might be possible for non-graduates who have developed their coding and other technical skills to a very high level to find employment.
You could either take a degree in AI or a related subject or keep your options open by studying another quantitative discipline (and then possibly specialising at postgraduate level). Suitable undergraduate degree subjects include data science, computer science, maths, statistics, physics and electronic engineering.
Entry requirements for degree courses vary, so check carefully with individual universities and through UCAS; many courses, for example, look for an A level/Higher in maths. An alternative qualification, such as a suitable BTEC Level 3 National or T level, may be acceptable, but do your research.
TIP: Employers often seek people with relevant experience so a placement, sandwich course or internship would be valuable. Find out what links universities have with employers and look out for programmes accredited by professional bodies.
Sometimes employers expect you to have taken a relevant master’s degree or even a PhD, or it might be possible to continue with your studies part time in employment. A growing number of postgraduate programmes are available in AI, machine learning and other relevant subjects.
TIP: Apart from gaining academic qualifications, take the initiative to develop your coding and programming skills. Igor says, “Try to implement everything you’ve learned. There’s a shortage of people who can write nice, clean code, so knowing how to do that will make you highly employable.”
Once working in machine learning, you’ll be provided with training for your particular role and your employer may support you in your career development. Some offer formal graduate training schemes or apprenticeships. In England, for example, there are degree apprenticeships at level 6 for data scientists and at level 7 for AI data specialists.
How did Jarosław and Igor enter this area of work? Igor explains, “My first degree was in mechatronics and I also studied robotics. My master’s thesis was about image inpainting. Alongside my studies I worked in machine learning support in an innovation team in industry. After I graduated, 18 months ago, I started working for Digica.”
Jarosław’s route was very different, which shows that there is no one set career path. “I’m actually a physicist. After gaining a PhD I spent about 15 years working as a scientist on various projects in academia. Six years ago, however, I decided to transition to a more commercial environment and secured a job in data mining at Infosys. After two years I moved to my current position at Digica.”
What opportunities are there to progress in machine learning?
Demand is high for those with expertise in machine learning, so career prospects are good both in the UK and overseas. Promotion to more senior roles normally involves managing a team.
With experience, you could work as a consultant. You could also consider becoming a self-employed contractor or freelancer, or even set up your own AI business. Other options include moving into another area of computing, working in research or teaching.
TIP: Joining a relevant professional body, such as the BCS, The Chartered Institute for IT, The Institution of Analysts and Programmers (IAP), The Institution of Engineering and Technology, the Alliance for Data Science Professionals or AISB (The Society for the Study of Artificial Intelligence and Simulation of Behaviour), will help with your continuing professional development. Most offer networking events, courses etc.
How can I find out more about careers in machine learning?
More information on AI and machine learning may be found through some of the professional bodies mentioned above. Career profiles are available through Prospects, My World of Work, Careers Wales and targetjobs.
Finally…
A career in machine learning will give you the chance to be involved in a rapidly developing area of work and the satisfaction of helping to provide solutions to real-life problems.
Igor says, “Despite the challenges, understanding how AI works is too satisfying to let go!” And Jarosław explains, “New technologies allow us to solve much more complex problems and automate many processes. It’s fascinating to consider that, perhaps one day, problems that were once unsolvable will find their solutions. If you’re interested, give it a go and don’t be afraid of change.”
Debbie Steel, December 2023
With a background working with apprentices and teaching in further education, Debbie was employed as an in-house careers author before establishing herself as a freelancer. As well as co-authoring numerous careers books, Debbie has produced resources and web content for a range of high-profile clients. She is an enthusiastic proponent of impartial and reliable careers information, and a member of the Careers Writers Association.
.png)
