The measurement might be legit. The comparison IEM seems to be Sony XBA-4, which sounds exactly as it measures and does have that 7k peak plus some ringing, similar to Sony MDR-EX1000.

Tyll's IEM compensation seems to understate 4-6k range and that IE800 measurement also suggests some lack of seal. Subbass drop is a dead giveaway - not to mention the coupler might be plain inaccurate there. Another is the IR - looks like a massively underdampened driver and would sound like junk.

BA drivers tend to have peaks in 2-8k range, at least a few smaller or larger. At least all Knowles and Sonion ones have. That's why some have holes drilled in their driver plates (type I, II, III damping) or have ferrofluid added - to dampen those resonances. Plus sometimes a filter of choice (e.g. Phonak's). Dampers tend to make these peaks stand out more instead for some reason.

ER-4S old Knowles armature sports the 3k resonance of the mild variety.

I found the 5k or 6k peak to be the common kind, BA or not. Probably related to how hard it is to have the right shape of 1-3k HRTF hump yet keep the rest below that.

Even 7mm or so drivers can lack subbass - it's all in how the driver was tuned. This one has too much venting most likely, which could be fixable.

I'm pretty sure a good crossover can have reasonable phase distortion. By that, I mean below 30 degrees in a non-critical range - the problematic part is crossing the BA supertweeters. That said, a 4 BA design should probably use 2 of them as supertweeters to extend FR up as much as possible, as well as one extra bass driver.
This Sony has a cross right in the range where hearing is sensitive.

For some reason, almost all BA IEMs have a problem reproducing anything > 16kHz. A few can go higher, but most face a brickwall, like that XBA-4. And no, it's not possible to equalize that - attempting to do so will overload the armature and cause damage or huge distortion. Doing this to dynamic drivers seems actually doable as long as the amplifier can drive higher inductive or capacitive (pick your poison) loads and remain stable.

Odd order active Butterworth or Chebyshev I low shelf with gain is actually best for BAs, as it corrects the usual positive phase. (The amp needs some phase margin in highs for that, as it will probably remain slightly negative.) Use even order for dynamics to distort the phase less in general.