Generally, those questions concern MTurk rather than Inquisit. I would thus encourage you to study MTurk's documentation for answers.
, we are not sure how the subject ID number should be assigned to assure the correct randomization of the in-built experimental conditions
#1:This entirely depends on how you set up the experiment in the first place. I.e., how many conditions are there? How do you plan to assign participants to those conditions? Via subject id or group id? Should they be randomly generated, subject-entered, passed by query parameters, etc.?
how to generate the MTurk code at the end of the study for Ps to enter on Amazon for compensation.
#2: Use Inquisit's <expressions>, <values> and the various available functions to generate a unique code and display it to subjects at the end of your Inquisit experiment. Have participants submit that code to MTurk. Match the codes between Inquisit and MTurk to establish "proof of work".
Cf. http://www.millisecond.com/forums/Topic6753.aspx#6753
Cf. http://www.tylerjohnburleigh.com/mturk/2013/01/verification-code-part-1/