In my experience, the design of these kinds of chokes is an exercise is optimization. Usually, there are constraints like size and temperature rise which govern what the final design needs to be.
Differential-mode (DM) chokes: Typically wound with alloy powder / iron powder core, low permeability (typically between 60u and 160u). "Sendust" and iron powder are perhaps the most common(?) / cheapest. But, something like "high flux" will yield a smaller inductor due to greater saturation flux density. The inductance of powder cores rolls of relatively gently with applied MMF (i.e. current). You might design this choke to have 30% roll-off in inductance at the peak current you expect. Though, whether that is acceptable really depends.
Core size and temperature rise are related because the core ID dictates the wire size you can use. You might typically fill ~30% of the ID of a toroid with wire. (total cross section of wire / πr^2 < 30%). I would not go higher than 50% fill.
If, to achieve 30-50% fill, you determine that you need wire which is too small for the rated current, then you should consider a larger core with a larger ID. Judging what size wire is big enough depends on the thermal resistance (to ambient) of the wound choke. This will vary depending on construction, size, and if there is any airflow. With static air, you might guess around 50-75°C/W for this size. So, you probably don't want too much more than a watt of dissipation.
Common-mode (CM) chokes: The process is similar. You need to pick a core size that allows you to fit the number of turns you require with wire of sufficiently large diameter. Because of the equal and opposite flux, you can use a much, much higher permeability core. Most commercial EMI filters will use between 7,000u - 12,000u ferrite cores for low frequency common-mode chokes.
CM chokes typically require more turns of wire to achieve adequate attenuation compared to a DM choke in the same filter. (At least that's my experience.) CM chokes may have comparatively more interwinding capacitance, or "self capacitance". You might need to consider the self-resonant frequency of the choke and whether it makes sense for the EMI you are trying to attenuate. You can measure this with an inexpensive nanoVNA.
For example, if your 33mH choke has a self-resonant frequency of 50KHz, and the EMI frequency is 500KHz, it won't do a very good job. This is to say: Inductance is an OK parameter to design with, but be wary that it is just a single point on the impedance vs. frequency plot.