Traditional hiring processes are very time-consuming, often lead to inconsistencies when the resumes are reviewed manually by HR , and is likely to make some mistakes. Many existing recruitment tools today use keyword matching, fail to capture the actual meaning of resumes limits fairness and there is no transparency in the candidate selection process. This proposed approach works as a TF-IDF based machine learning model to evaluate how well a candidate’s profile match with the job requirements. Preprocessing involves removing the personal information from resumes , followed by normalization , tokenization. Conversion into TF-IDF representing the relative importance of skills and experiences. The trained classifier model , developed by using a dataset of 350 Kaggle resumes and has a strong predictive performance with the high accuracy 98.08%, 97.5% in the precision, 96.8% in the recall, and an F1-score of 97.15%. SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) are employed to explain how the model makes its predictions, ensuring transparency and fairness. This makes the model is explainable and that bias is found. The system enables real-time inference by rapidly processing uploaded resumes and producing prediction and explanation results without manual latency.