Given the current economic climate where large companies are laying off machine learning employees in large numbers, you may be wondering if it’s worth spending four years and over $80,000 on education. If you competed against hundreds of candidates for the few positions posted, how long would it take you to get the job? What salary could you expect?
These days, many machine learning jobs in the US offer salaries well below $100,000 a year, especially for beginners. Many still offer $200,000 or more, but they usually require professional experience and are not usually something you learned in school. It raises more questions. There are multiple sides and ways to this.
Degree type
Some people get a degree in a field of their choice, not for a job. Climbing Everest requires a lot of training. You may spend a lot of money to achieve your goals, your passions, but you won’t get any money in return. PhD is the same. We all know that no matter how good you are, getting a PhD in academia is very unlikely to get you a decent salary and tenure. Yet the degree is designed for that very purpose and offered because it is in demand. In fact, having a PhD can hurt you in the job market and some applicants don’t include it in their resumes. People do it for other reasons: recognized prestige, Potential credibility and visibility (if you write a book), and passion for research.
But what about a hands-on master’s degree that includes internships, real programming, portfolio building, and all to impress potential employers? Great timing for those just starting out. First, given the shortage of jobs, it’s more productive to spend time studying than looking for hard-to-find jobs these days. But the economy of machine learning professionals will recover. I see more and more recruiters looking to hire even though I only hear about layoffs. Some of the dismissed do not have a real degree, but have a certificate or data his camp. There are good providers, but there are also many bad providers (who usually won’t even tell you who the instructor is). Recruiters these days are more likely to require an actual degree.
What companies usually do
Some employees were overpaid or failed to sell the value they created to their bosses and decision makers. Companies will not cut salaries to adapt to new markets. Cut jobs while hiring new (cheaper) recruits and boosting salaries for your best employees. It periodically causes or occurs because of recessions. Those looking to enter this market with a good education and reasonable salary expectations will eventually win.
Impact of ChatGPT
Will AI replace workers in the future, including the people who develop it? There is no doubt that AI can do many things, but there is also a lot of hype surrounding it. It’s easy to imagine that many time-consuming and tedious tasks such as debugging and data cleaning will become increasingly automated, but we’re still far from there. Even basic problems like scoring news from fake to real still need to be solved. It can be solved without learning algorithms (more on that in a future article), but for now, many people working on this (machine learning engineers included) are using these models to I’m not good at using algorithms, let alone training my own brain. Some companies prefer fake news because it is clickbait and a source of revenue, but they will face competition from companies tackling this problem.
Hype, and what stays here
Speaking of hype, vendors try to sell expensive products like GANs and deep networks and find buyers because of the hype. I had to include it in the class I provide as well. Because that’s what many participants (machine learning professionals working for insurance, health, or financial companies) want to hear. Ultimately, buyers will find that for relatively simple needs, cheaper solutions may work just as well. The market is adjusting accordingly.
But the bigger long-term problem that will affect all jobs, not just data science, is declining population growth, which will eventually turn negative. This will open up more and more positions in the healthcare and related sectors serving the elderly. Addressing climate issues is unlikely to run out of steam anytime soon.
About the author
Vincent Granville is a pioneering data scientist, machine learning expert, founder of MLTechniques.com, co-founder of Data Science Central (acquired by TechTarget in 2020) and former VC-funded executive, author, and patent owner. Vincent’s past corporate experience includes Visa, Wells Fargo, eBay, NBC, Microsoft, CNET and InfoSpace. Vincent is also a former postdoc at the University of Cambridge and the National Institute of Statistical Sciences (NISS).
Posted in Vincent number theory journal, Journal of the Royal Statistical Society (Series B), and IEEE Transactions on Pattern Analysis and Machine IntelligenceHe is also the author of “Intuitive Machine Learning and Explainable AI”, available here. He lives in Washington state and enjoys researching stochastic processes, dynamical systems, experimental mathematics, and probabilistic number theory.
