THE HIGHEST STANDARD IN AI — AI GUILD #DATACAREER
The Golden Age: Data Careers in the 2020s.
Take charge of your career development to Lead, Director, and Chief Data Officer.
For 2020, we have global data showing half a million professionals in Data Science and Machine Learning. By comparison, the number of software engineers was about 15 million, or 30 times more.
A zettabyte is now the unit for messing data volume growth. A zettabyte is a trillion gigabytes. From 2020 to 2025, the global data volume will grow threefold to reach 200 trillion gigabytes (200 zettabytes). Hence, 10X is a plausible scenario for the profession’s growth in the 2020s.
Ten years have passed since the Data Scientist was declared the sexiest job of the 21st century (2012). Not always has life matched expectations. Many early professionals were working in small teams or alone. Companies got stuck in the proof-of-concept cycle. It was reported that only 1 in 5 use cases made it to production.
Ten trends shaping your data career to 2030
Given the ubiquity of data, the rise of large models, and a decade of experience, I want to show that the golden age of the profession is now ahead of us. Over the next ten weeks, I will survey the landscape and argue the following ten hypotheses.
- There is no lack of talent.
- It’s about the data, not the algorithms.
- Digitization is a bad choice.
- You need use cases in production.
- Three people are not a data team.
- You need an infrastructure for multi-modal data processing.
- Be the go-to expert: Data roles are evolving by industry and function.
- Your company must embrace data literacy.
- Getting use cases in production is a hard problem.
- Businesses and regulators need leaders they trust.
To read more, join the discussion, and, make your contribution, please follow Fluent in Data to receive your weekly update.
The ten trends previewed
There is no lack of talent.
Graduates, including PhDs, often have years of experience with Python and data. All professionals can learn the basics on the Internet. I believe there is enough talent now, and from my work, I would say that sometimes the talent is already overqualified.
It’s about the data.
Algorithms alone won’t do it, and startups selling algorithms are not flourishing. The rise of the large models reinforces the insight that it is all about the data. I expect that the biggest challenge for companies is to consistently leverage their data and connect it to large models so that use cases can be productionized profitably.
Digitization is a bad choice.
That software is eating the world is a bonmot, but it is not well understood that software only consumes the world successfully if it generates and captures data. Therefore the notion of digitization or digital transformation is a bad choice because it leads businesses and governments to focus on software development only. Thus they miss joining the data economy.
You need use cases in production.
If you are not part of a team delivering use cases in production, then your company needs a strategy for productionization now. If it does not have one, you are better off moving to a company with use cases in production. Use cases are where you will make a career in the next ten years.
Three people are not a data team.
The size of the team matters. It is a good proxy for whether the company leadership is data literate. Data requires investment, and first, this must be in-house to build and leverage the data infrastructure. The overall data team should have dozens, if not hundreds, of members working on one integrated production system. Data is not a project; it is a product. If your company does not get it, then find one that does.
You need an infrastructure for multi-modal data processing.
It sounds technical and points to two fundamental quality issues. If your data is corrupted at source, the pipeline and model investment will unlikely pay off. Also, you want your pipelines to serve as many use cases as possible. For that, you need engineers, and you must lead them from the perspective of the customers and users.
Be the go-to expert.
Data roles are evolving by industry and function. We understand the difference between a Data Scientist, NLP engineer, or MLOps specialist. And more roles will emerge and become specialized. However, the specialization is not just technical. Increasingly you need relevant domain expertise, e.g., the industry, customer, and particular data type. Indeed, the path to leadership is creating value from data, and for that, you need a career path that combines technical leadership with domain expertise.
Your company must embrace data literacy.
The literacy target is 100%, always. Literacy means that all can read, write, and do numbers. Data literacy would be a company’s ambition that all employees understand data, use data, and contribute to data collection, generation, and processing. To my mind, only a few companies worldwide achieve this. Too many companies fail at the executive level because the leaders are not familiar or comfortable with data, statistics, and probability. To my mind, data literacy is the competitive advantage of the 21st century.
Getting use cases in production is a hard problem.
Large models, cloud deployment, and no code solutions: Sometimes you hear that the Data Scientist will be superfluous. As I am writing from Germany, please consider the following analogy with the automotive industry: Getting a Machine Learning proof-of-concept into production is no less simple than getting a concept car into production.
Businesses and regulators need leaders they trust.
The chief data officer in any company urgently needs a seat at the table. To date, the CDO, too often, is one or two levels removed from the inner circle, reporting to someone else. Yet, the data profession also needs to do quality assurance urgently. Software engineers have avoided professional accreditation successfully. I argue that data professionals are more like engineers and doctors, and professional accreditation is required, perhaps unavoidable.
The AI Guild’s 2000+ Specialists
The AI Guild is Europe’s leading practitioner community in Data Analytics, Data Engineering, Data Science, Machine Learning, Deep Learning, NLP, Computer Vision, and MLOps.
Do you want to progress to Senior, Lead, and Director?
It takes 60 days to build your competency profile. You can find out more by booking the first conversation to gain more insights at https://www.datacareer.eu.
