Hi, iam a python expert i have experience in fine tuning chatgpt, it is very easy. The two aspects which are very important is the "prompt and completion" pair which is the question and Answer pair in leyman terms which is the input given to the chatgpt while training. Meaning the data should be kept in that prompt and completion pair manner for training it. And depending upon on the specificity of the domain the training time and cost will depend. Usually chatgpt learns quicker as it is a pretrained LLM when compared to others, so yeah let me know if you want me to work on this project.