Some of the pros<p>- good pay<p>- exciting time to be alive<p>- work often spans broad business departments, involves genuinely interesting engineering trade-off problems between accuracy, scale, budget, and interpretability<p>Some of the cons:<p>- You have to be wary to avoid companies in which machine learning is just a name for data engineering / devops / business analytics, or where it is hype or aspiration and no substance<p>- No matter what your job will mostly be the unsexy stuff to get a system working. Only maybe 10-20% is going to be the actual model development or experimental side<p>- Companies not yet well versed in machine learning can have huge sticker shock at what a competitive compensation package is like, and waste a lot of your time on interviews and end up way off the mark on compensation.