How do I get the "Ideal Characteristics" of a candidate for least attrition in Machine Learning?

I am working on a project to predict whether a candidate, after joining our organization, would leave us within 1 year or not. The model is based on different features present in their resumes (average tenure, skills, degree, etc.) and features of our organization (job role, place, supervisor, etc.)

The aim is to get the ideal characteristics of a candidate who won't leave within a year. This information can be then handed over to HR for better recruitment in the future.

So far, I have applied Chi2 test to filter out some categorical features with no significance. I am using a random forest model for its interpretability. I have tried Recursive Feature Eliminiation on continuous features, and SelectKBest using Chi2 on categorical features to keep the number of features low. I am able to get the feature importance as well using the Random Forest.

But I am not able to get what I am actually interested in, i.e. ideal characteristics or the top sets of values for these selected features.

Please help me with how I can proceed with it. I have tried on Google too without any luck.