I have posted a similar question to this previously, but would like to reword it and re-ask it.
I just recently completed my undergraduate studies with a degree in mathematics from a relatively well-known US university. During my time as an undergrad I took many advanced (first and second year graduate courses) as well as one-one reading courses with professors on topics in Probability and mathematical analysis (topics like stochastic differential equations, large deviations theory, linear/nonlinear PDE at the level of Caffarelli and Cabre, as well as topics in Calderon-Zygmund and Littlewood-Paley theory).
Due to my indecisiveness and the insistence of my family for me to find "gainful employment", I chose to not apply to graduate school, and instead work full-time in private-industry (quantitative-finance). The work is challenging and sufficiently interesting, and of course pays well, but I still feel that I would be remiss if I did not purse a PhD in mathematics (or related fields). My main issue (in my mind) is letters of recommendation. While I am confident that the professors I have worked with would be willing to recommend me, I am in some sense taking a non-traditional route by delaying my application (for example after one year of working). In turn, I am afraid that the more time I take for asking a recommendation, the less-likely it is that my professors will remember me and be able to write a strong letter for me.
Does anyone have recommendations as to a best course of action here? I am wondering whether I need to bite the bullet and apply to programs this year, and hope they offer deferrals (for say, one year of time) although I'm not sure how the current situation (with covid etc.) could impact this. If anyone here is on mathematics graduate decision boards (or any graduate decision boards for that matter), I would appreciate your input. I am also wary as to the perception of "non-traditional" applicants as my self and how they might be perceived by admissions committees, although it is too late to modify that status.