This post is adapted from my answer on Reddit.
I think there are various career paths that fall under “Data Science/Machine Learning”. This doesn’t mean there are 4 roles, it’s just how I see this industry; nevertheless, the more skills you possess, the better. Below are my personal thoughts based on my limited knowledge, some may be outright wrong or inaccurate, so take with a grain of salt.
Data Analyst/Business Analyst - need to understand domain knowledge and be prepared to do lots of research on both qualitative and quantitative data. They will also need to answer business questions both internally and externally. For example, whether some analysis is feasible given the data we have. They will focus more on the story-telling side. There is some overlap with (management) consultants in our data-driven world. Their stakeholder is probably directors/managers from various departments that have their data analyzed by these analysts.
Data Engineer- need to understand the underlying infrastructure (DB, pipelines, CI/CD, distributed systems) to build efficient and scalable systems to store/process data. Their users are both data analysts and machine learning engineers.
Machine Learning Engineers - need to understand and apply various AI models and approaches. They need to know “when” to apply “what” to “which” problems. Their users are the data/business analysts.
When - the domain context (E.g. healthcare, finance)
What - the AI/ML models or approaches (E.g. convolutional neural networks, support vector machines, genetic algorithms)
Which - the actual problems, which is usually comprised of a question or hypothesis
Research Scientists - I personally only call people who do research as scientists, whereas people who only apply (or piece together) what’s available as engineers. Both groups create lots of value in the world, we can’t live without either. These people try to make progress in visuomotor problems in AI and robotics by developing new models and architectures to improve the agent’s capabilities such as image recognition, natural language understanding, decision-making, reasoning or even control theory in robotics.
Anyways, nowadays for jobs, the buzz is too much for job titles to actually reflect the responsibilities of the role. So always look under the hood on what types of problems the company is solving and why they’re trying to hire people.