khushal patil

I’m an AI/ML Engineer and full-stack developer specializing in Python, ML, and the MERN stack. I build practical systems that detect risk in real time—whether it’s fraud patterns in transactions or scams hidden in job postings.

In my projects, I developed an NLP-based fake job posting detection platform and a real-time AI banking fraud detection system. The fraud platform evaluates transactions within 100ms using an XGBoost model trained on 284,807 records and achieved an AUC of 0.9792 with 89% recall for fraud.

I also focus on performance and explainability. I tuned models for imbalanced data (using SMOTE), compared multiple algorithms to select the best performer (e.g., XGBoost), and integrated SHAP to provide clear reasons for every flagged transaction.

I’m co-authored a published ML research paper and enjoy building end-to-end ML pipelines—from preprocessing and feature engineering to training and performance visualization. I’m now seeking a machine learning / software engineering internship to apply my ML expertise to real product development.