Rupesh Roshan
@rupeshroshan
Associate Developer at NPCI
Hyderabad, India
Rupesh Roshan is an Associate Developer with experience in building advanced fraud detection models for the National Payment Corporation of India (NPCI). He has built several ML models for P2P and P2M fraud detection, serving almost 18 crore UPI transactions daily. His expertise includes feature engineering, deep learning, and developing data pipelines using Python, SQL, and various ML frameworks.
Experience
Associate Developer
National Payment Corporation of India (NPCI)
Responsibility includes fraud Model building, feature engineering, Exploratory data Analysis, Model training and testing, data pipelines script generation, scheduling, data docs preparations. Built several ML models for P2P(Person to Person), P2M(Person to merchant) fraud detection which serves almost 18cr UPI transactions per day in production environment. Contributed in Generation of more than 200 features for every customer on UPI platform, EDA, Visualization, automating training and testing of ml models. Built vpa (text) clustering model using NLP library. Contributed in building the architecture of International UPI. Tech: Python, Machine learning, Deep learning, NLP, SQL, MLflow, Big Data, Cassandra, Postgresql.
Education
Indian Institute of Technology (ISM) Dhanbad
M Tech
Worked with Prof Ansuman Bhattacharya to create predictive model for An Efficient Video Surveillance using Deep Learning A real time arms detection from captured video by robot in specific situations like (cordon and search) using deep learning (CNN)achieved accuracy of 93.2%. Tech - Convolutional Neural Network, Keras, Python, Image Processing, Matplotlib, Transfer Learning.
G L Bajaj Institute of Technology and Management
B Tech
Suraj Bhan DAV Public School
Senior Secondary
Mithila Public School
Higher Secondary
Licenses & Certifications
Data Scientist with Python
Apache Cassandra Developer Associate Certification