I would definitely say so. In my last 3 jobs as both a data science manager and data scientist, I’ve tried to use machine learning but in each job it is the same story.
The company states in the job description they want somebody with ML knowledge and the know-how to apply ML to solve current problems. Once you get hired, it is a completely different story.
- You find out you were hired to do just data analysis in excel or R or Python because no ML currently exists.
- You find out data is sparse and that the company actually doesn’t have the right features to do what it wants. Ultimately, real-world problems show a high degree of non-linearity and the easy canned ML solutions that you learned in college don’t really apply. A typical real-world scenario is, you run a model and the model comes back with a 45% accuracy, which is not much better than a 25% random guess. At this point, flipping a coin would be better than machine learning.. lol
- You find out you are a SQL junkie and you write SQL all day long to retrieve data. 90% of the hype is around getting the data right and building dashboards. lol
- You find out you have a manager who is apprehensive about ML because he feels it is a black box and he doesn’t know what is going on inside it. So he is not about to let ML take over even as an experiment because he feels ML is unpredictable and that the business may lose money. So your dreams of putting ML models into production are a pipe dream far in the distant future.
- You find out IT is not cooperative and that they see you as the hotshot new kid in town with skills that nobody understands. They are afraid that if they let you take over it will diminish their power. Some simply don’t understand and they feel you might take over their jobs. For example, ML has largely moved to the cloud now but in order to access those cloud resources and serverless architecture to implement ML in production, you need full rights. IT will simply deny those rights to you because of the aforementioned reasons. The manager doesn’t want to upset the current working conditions, so he won’t know what to do because the bottom line is to keep the current business running and the priority is not to have you try fancy ML techniques that are unproven which may or may not bring a profit.
Ultimately, you are lured into the job with aspirations of putting big ML models into production and doing cutting edge algorithm research only to find out that the company is not ML ready and that they only hired you so they can say to their clients, “Yes, we have a machine learning guru onboard”.
You can always try to start your online business and with the proper web hosting partner, your success is guaranteed.