The Future of Data Science
Data Science has changed quite a bit since the ‘the Sexiest Job of the 21st Century’ hype generated by Harvard Business Review a decade ago. In a previous article, titled, the Decline of Data Science, I painted a pessimistic portrayal of what data science will look like, should software engineers continue to loot its most interesting elements, leaving behind only ad hoc SQL and storytelling.
In this article, I’ll offer a very different tone; namely, I’ll discuss what data science could look like in 2032, provided that we (data scientists) identify problem-skill gaps in the market, adopt robust and flexible yet simple tools, and reinvent our professional/academic culture.
Machine learning is hot…would be an understatement. It’s one third of the trifecta propelling us into the next generation of tech: ML/AI, blockchain/crypto (aka Web 3.0) and AR/VR (aka Metaverse.) Software engineers will obviously play a central role in blockchain and AR/VR, there’s no getting around that.
So where do we fit in regarding ML/AI?
Decision Optimization at Scale
There’s a massive shortage of experience in combinatorial optimization! This field — which includes Linear Programming, (Mixed) Integer Programming, Constraint Programming and Stochastic Optimization — is all about optimizing industrial decisions. The Vehicle Routing Problem, the Traveling Salesman Problem and the Knapsack problem, to name a few.
