Misses thin characters

[khilaadi01]

When using this gem to read PDF resumes, I'm noticing that it misses some characters consistently. For example it reads 'Netflix' as 'Netix' or 'five' as 've'. When I read the same pdf file using the 'pdftotext' tool (from Poppler) in my local machine, it reads the words just fine.
The characters in the PDF seem to be embedded correctly:

name type encoding emb sub uni object ID

---

AAAAAA+Lora-Regular CID TrueType Identity-H yes yes yes 17 0
AAAAAA+Lora-Regular CID TrueType Identity-H yes yes yes 7 0

[jbigler]

fl and fi are two common ligatures where the upper loop of the f joins with the next character.