Suppose in general, if I use AWS Lambda layers for dependencies like pandas or tabula-py, which individually can exceed 50+ MB, do I need to create a separate Lambda layer for each dependency if my project has around 10 such libraries?

I'm trying to understand the best practice here:

Should I bundle all heavy dependencies into one layer?

Or should I split them into multiple layers, one per library?

Also, how do I handle size limits in this scenario?

Explored Lambda layers, but not sure about the layer strategy when multiple large libraries are involved.

What i expect -

A best-practice recommendation for managing multiple large Python dependencies using Lambda layers