ML Job Roles

It takes a multi-disciplinary team to build an ML product. Yet, as the entire AI field is very new and vast, job roles & titles aren’t well standardised between companies. Job titles and responsibilities can vary from place to place, with overlap and nuance which easily gets lost in translation. At many of the companies I’ve seen, ML teams have contained some combination of the following roles.

ML Scientist

Machine Learning Scientists are the people who understand the ins-and-outs of ML algorithms. They understand which algorithms to use, can build, evaluate & deploy your company’s ML models, know how the data is impacting performance, and have ideas for evolving the ML technology to give better performance over time.

Data Scientist

Data Scientists are closely related to ML Scientists, and they tend to work more directly on analysing and modelling data to provide actionable insights for your business.

Software Engineer

A software engineer will design, write, test & deploy the code behind your product, then maintain it as time goes on. They’ll fix bugs and add new features as appropriate.

MLOps Engineer

MLOps is a relatively new idea, and so MLOps Engineer is an emerging role. The parallel comes from DevOps Engineers, who deploy and maintain production code. MLOps Engineers deploy and maintain production ML models, building processes and tools for making these tasks more systematic.

Data Engineer

Data Engineering is closely related to Software Engineering and MLOps. The job of a Data Engineer is to design and build infrastructure to store data, and pipelines for processing it, which are crucial aspects of any ML product.

ML Product Owner

A product owner is responsible for the product roadmap — understanding both user needs and business requirements, and prioritising features effectively to build a product that brings value to your business.

Machine Learning Manager

The job of an ML Manager is to keep your ML team performing effectively, looking after their work, their careers, and the general health of the team. They’ll usually be a technical specialist themselves, and able to smooth communication between the technical & commercial parts of your organisation.

Computational Linguist

This one is quite specific to voice and language technology! Knowledge of linguistics & management of language artefacts (like pronunciation lexicons, grammars, text normalisation rules etc.) is often crucial to getting voice and language technology to work effectively.

UX Designer

There’s always someone using your product, and the way they use it is designed by a UX (user experience) or a UI (user interface) designer. The latter is often focused on a graphical interfaces (GUIs), but there are other modes of interaction like voice where you might find other roles such as conversational designer.

Data Annotator

To build ML requires data, and that data must be labelled accurately, usually manually. While some companies build this function in-house, there are many third party companies that specialise in data labelling.

Domain Expert

What a domain expert is will vary from application to application, so this last bucket is a catch-all for everyone else who contributes! If you’re building a medical application, then medical expertise from doctors or other healthcare professionals can be invaluable. On the other hand, a legal application may need insight from lawyers. The key thing is to ensure that your ML team have access to the right domain expertise to build a valuable product.

In practice, while the roles are listed separately here, people might take on multiple of these roles depending on the size and shape of the company and their individual skillset. This is especially the case in a small company or in an early-stage product. For example, a software engineer’s job might include data engineering, an ML scientist might have a big part to play in product ownership, or a computational linguist might also do data annotation. But, as you can see, a successful ML product is built not just by ML scientists, but by a host of different people with different roles on the team. Getting the right balance is one key part of building a successful AI company.