Data scientists are disappearing. No, not in the physical sense (no rapture here), but in the job market. The term “data science” has been a catch-all term for years, but as companies better learn what goes into hiring data science teams, the generic “data science” job titles might go the way of the dodo.
Not that we’re sad about it. Confusing job postings and a misunderstanding of the fundamentals of working in data science has led to disillusionment on both sides. Stakeholders are frustrated that they aren’t getting results, and data scientists are frustrated that their results aren’t acknowledged. Change is coming.
In 2020, we’ll see a continuation of the specialization process with more companies getting specific about what they’re looking for as a way to get the results businesses and organizations need. Here are some data science job titles that are emerging or continuing on their popular and merry way.
You saw that coming. Data engineers provide the backbone and the structure for all sorts of data scientists to do their jobs. Here’s the thing: data engineering helps translate pure data science into something that creates business value.
Data engineers are responsible for the pipeline for data and its products. They’re responsible for making the technical decisions about frameworks and environments so that data scientists can get creative. When something breaks down, they’re responsible for finding out what the problem is.
For every one data scientist out there, a team of multiple (some predict up to five) data engineers continue to monitor, maintain, and troubleshoot the environment. They can direct data science efforts and build those business-worthy pipelines. And with almost 131,000 job postings on Indeed alone, versus just over 12,000 data scientist postings, it seems the tides are turning.
With so many new products on the market requiring AI processing, companies are looking to manage the complex aspects of this new “team member” with an expert. Product management arose around the 1980s as a way to help keep both the historical values of the company at the forefront and to ensure the pipeline ran smoothly.
AI Product Managers take those management principles and apply them to AI-driven initiatives. They define the AI problem and compile the data to train models. With AI-bias an increasing problem and black box solutions still an issue, AI Product Managers can help define what it means to build these initiatives.
Right now, it’s tough to hire an AI Product Manager because the job title is still relatively new (and may not even be called that in some postings). Companies willing to hire someone with leadership skills, experience in product management, and the technical understanding of things like supervised and unsupervised learning can create the position by honing in on talent.
Businesses need these Product Managers to help with consistency before, during, and after AI adoption. AI Product Managers are also responsible for the human side of AI, oversight, synthesis, and critical thinking.
A partner to the AI Product Manager, the AI Architect takes the problem framework and the directives and builds the environment for the solution. The Data Architect is the leader of the data science team, helping to understand the technical implementation of the project and troubleshooting when the pipeline goes south.
AI Architects are practical. They have the technical and theoretical knowledge of pure data science with business training in the mix. They understand the practical implementation of the initiative and have the business sense to handle issues that arise from the business aspect.
The Data Architect frees up the data science team to work through models without having to figure out the technical aspects of why they aren’t working and decides on the technology to implement the initiative.
Data Architects should be able to make judgment calls based on the conflicting needs of the business side and the idealism of the data science team. Limitations and restrictions specific to the business are also on the table. It’s a highly specific role that does require some level of experience and a natural sense of problem-solving.
Data Analysts, this is your time. Data is messy, and it mirrors our own biases and beliefs about the world. The value of a data analyst goes beyond just cleaning data and extends into making judgments about how best to approach the messy world of data for clean, reliable results.
Data analysts help build security into the data process. A good data analyst can analyze the data and scrub it so that it’s ready for processing. That same data also needs to be accurate representations for model training, and if bias is detected, data analysts should be able to examine that data for potential issues.
If AI is a black box solution, that information analysis is even more critical. An intimate understanding of what’s going into these models will be the only avenue for fixing a flawed model until truly explainable AI is available.
Cybersecurity isn’t typically in the realm of data science. Still, with the emergence of AI-driven initiatives, information security analysts will need to have a clear understanding of the functions of data science to be prepared. Blending a knowledge of data science with cybersecurity will be a new role for data scientists to fill as this tech progresses.
Information security analysts understand risks based on context. The dance between data availability and security is delicate. This role takes the principles of AI, machine learning, and some behavioral analytics to provide the best path for security in the age of artificial and augmented intelligence and big data.
Data science job titles that fall under the data science umbrella are still a sure fit coming into 2020. As we build our understanding of new AI-driven business initiatives, the aspects of data science become even more critical. Niche into one of these categories, and you could find yourself landing a rewarding career.