Depends on the job.
I use a lot of computational analytics, which is largely building out models using software with proprietary physics, but the modelling is mostly huge amounts of degrees of freedom (so controlled with very large matrices) and the physics is the governing equations of fluid mechanics, thermodynamics or kinematics ran over time-based simulations (basically a lot of calculus & linear algebra).
With the maths I don’t get my hands dirty as it were as the number of functions being solved is literally many thousands per second so it’s simply not possible to hand calculate but it’s important to understand what the model is solving, before starting a new set of simulations I may spend a significant amount of time going through the science looking at the coefficients & variables without numbers and making sure I fully understand what is happening (and I also have to spend a lot of time validating results).
Aside from this i do a lot of what i call fag packet maths, basically just taking odd equations to get some quick sense of scale for ideas in meetings in my case a lot of general data analysis mixed with thermodynamics (i tend to just do it on excel or matlab).
Some engineers really do very little maths that frankly is not much more then GCSEs.
In general your not going to be sat on a desk with paper & pen like you do when your learning physics & maths, but you do have to be very comfortable with what it is (& handling exceptionally large data sets with care), and once you’ve got it all you’ll be expected to summarise many weeks worth of simulations into 1 or 2 graphics on a couple PowerPoint slide.