Looking for a job

Raviteja visarapu

@ravitejavisarapu

Data Scientist at Ikspidia who redesigned fraud detection to improve recall from 5% to 40%.

At Ikspidia, I redesigned a real-time fraud detection system, improving recall from 5% to 40%. I deployed machine learning models that process 25 lakh+ transactions per day with less than 20ms latency.

I also built Kubeflow pipelines to automate experimentation and model evaluation, and used behavioral features and anomaly detection to improve fraud risk analytics. Previously, as a Data Scientist Trainee at Intellipaat, I worked on merchant segmentation, fraud models, and transaction data analysis.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Ikspidia

Mar 2024 - Present (2 years 8 months)

Redesigned fraud detection system improving recall from 5% to 40% on real-time transactions. Deployed ML models processing 25 lakh+ transactions daily with low latency.

Fraud Detection Random Forest XGBoost PyTorch Kubeflow DBSCAN Isolation Forest Autoencoder SHAP Feature Engineering ROC AUC Python

Data Scientist Trainee

Intellipaat

Mar 2023 - Feb 2024 (11 months)

Conducted merchant segmentation using unsupervised learning and built supervised models for fraud detection. Applied statistical testing and monitored model drift.

K Means DBSCAN Isolation Forest Logistic Regression Random Forest RNN Feature Engineering PSI Hypothesis Testing Python

Intern

Intellipaat

Performed data preprocessing and implemented ML algorithms across regression, classification, clustering, and NLP tasks. Assisted in feature engineering and model evaluation.

data preprocessing Machine Learning regression Classification Clustering NLP Feature Engineering Model Evaluation