I suspect, but cannot confirm, that this is a simple method that creates fonts according to the rules in the PDF specification (32000-1). It's therefore (probably) your job to deliver what 32000-1 needs, which is not necessarily easy or simple.
1. While widths are theoretically present in a PFB file, getting them is a beast of a job, requiring a full type 1 font parser. For this reason (and some others) PFM files were invented, which have simple binary tables of fonts. You would need the PFM file, but PFB and PFM usually go around in pairs. And you would need the PFM spec. Alternatively if the font is installed in Windows there are font APIs.
2. You cannot drop a PFB file into PostScript, it will error. I don't recall what has to be done, but it's something to do with removing the length fields which you (presumably) are already parsing. You need the PFB format specification.
I suspect you may also be missing a great slew of other information which goes into the Font Descriptor, and belongs in PDEFontAttrs, very little of which is optional (according to 32000-1).
It is almost certainly much easier to embed a font present in the operating system, using different APIs.