Minit Chitroda

I’m an entry-level data analyst and ML developer, currently completing my B.Tech in Computer Science and Engineering (Data Science). I focus on turning messy real-world data into reliable signals—whether it’s fraud patterns or audio authenticity.

In my Hybrid Data Analyst Intern role at Dcyber TechLab Pvt. Ltd., I built an end-to-end ML pipeline to detect fraudulent UPI transactions using Python, SQL, Pandas, and NumPy. I processed 6M+ financial records, engineered features, handled class imbalance with SMOTE, and trained models including Random Forest, XGBoost, and LightGBM.

I also delivered practical, decision-ready outputs: I built a real-time fraud alert system and created Power BI dashboards for monitoring and reporting. In a separate Remote Python Development internship at Debug With Digital, I strengthened my API-first development skills by building a conversational AI chatbot using the OpenAI API.

My projects reflect a deep interest in end-to-end modeling and deployment. For example, my Audio Deepfake Detection work used PyTorch with an ECAPA-TDNN classifier and a denoising pipeline inspired by RNNoise/Singly-Stage filtering approaches, achieving 94.7% accuracy and 93.4% F1-score on benchmark datasets, and I packaged it into an interactive Gradio app for real-time classification.