Integrate LLM to extract data from a PDF/DOC and prefill editor fields

[dmitrizagidulin]

The LLM is to extract data from the resume
send narrative chunks to the service to extract claims (this has a delay, note, its not fast); make sure its forgiving if the service is not available.

See also issue #49

[dmitrizagidulin]

Golda did a preliminary sample implementation of an API endpoint that does this, needs tuning and further implementation.