I was this close to picking up the PhD route in applied mathematics for science and engineering. I took up a job in a map company, "just for the time being" and not commit to the "Finite Element Method" or something around "Port Hamiltonian Systems" for 4-5 years straight, without knowing if it will find any applications in the real world.

8 years on, I'm still in the same map company along with 3-4 others who have joined after or mid-way their PhDs (and some more who didn't bother) and we do routing for hundreds of customers and basically have everything a PhD group has, including Seasonal Seminars on Routing Algorithms.

Finite element Method is not as exciting as AI (I used to think at some point that it would) and Port Hamiltonian Systems is something not a lot of people talk about, except maybe Dr. Volker Mehrmann and his group :)

Academia is not for the faint hearted and a PhD isn't just about research anymore. It's just a low paying job now, sadly. I still say things like "if I ever happen to go back to academia", as if I did a lot there. But I am an academic at heart, just that pursuing a PhD didn't certify it any better or seem to provide more freedom for exploring my passion for studies or a particular topic.