OpenAI GPT : Is it possible to set the context of the task with finetuning, instead of providing it as input for every API call?

Question

I am trying to create a GPT that can grade some essays. I am currently using the OpenAI API to input :

1. Essay Question,
2. Rules and concepts to look for,
3. Grading Rubric,
4. Essay Submission

every single time to grade an essay. I believe I am exceeding the input token threshold in every call, also this is going to be expensive in the long run.

Is there a way to feed/input the Essay Question, Rules and Concepts, and Grading Rubric one time using pre training or any other techniques, and reuse the API by just passing the essay submission?