Murali Krishna Bezawada — Data Scientist

Impact & Metrics

Numbers That Speak

Measurable results delivered across healthcare, fintech, and logistics — each number represents real business value.

Improvement

Claims AccuracyClassification accuracy via Random Forest & Gradient Boosting at Athenahealth

Reduced

Default Rate DropDelinquency reduced from 5% → 2% at Elevance Health via predictive behavioral analytics

Increase

Model Recall RateAchieved at American Express via ensemble methods and SMOTE resampling

Uplift

Customer SatisfactionThrough advanced segmentation & targeting at American Express

🏥

Patient Risk Stratification — Elevance Health

Deployed forecasting models and predictive algorithms for high-risk patient identification using Python, Scikit-learn, and Spark. Integrated LLM-based NLP pipelines via Hugging Face for enhanced patient risk assessment.

+15% fraud detection accuracy

📋

Claims Classification Model — Athenahealth

Improved claims classification accuracy by 20% using clustering and classification models with Scikit-learn and Spark MLlib. Built ARIMA/ETS time-series forecasting models to improve financial forecasting for healthcare providers.

+20% classification accuracy

💳

Default Risk Reduction — American Express

Reduced default rates from 6% to 3% via Logistic Regression, Gradient Boosting, and ensemble methods. Cross-product adoption grew from 40% to 60% using predictive behavioral modeling and A/B testing frameworks.

Default rate halved (6% → 3%)

🚚

Demand Forecasting & Pipeline — DHL

Built PySpark ETL pipelines for demand forecasting and pricing optimization across millions of logistics records. Applied SMOTE, PCA, and feature engineering to improve forecasting accuracy for supply-chain planning.

Optimized discount allocation efficiency

Career History

Work Experience

Data Scientist

Mar 2025 – Present

Elevance Health · Waukesha, WI

Deployed forecasting models and predictive algorithms for patient risk stratification, fraud detection, and care optimization using Python, Scikit-learn, and Spark.
Improved diagnostic accuracy by 10% through deep learning models (PyTorch, TensorFlow, Keras) applied to medical image analysis.
Enhanced fraud detection accuracy by 15% using computer vision techniques on healthcare documentation.
Reduced default/delinquency rates from 5% to 2% through advanced predictive modeling and behavioral analytics.
Built and tuned multi-layered neural networks and LLM-based architectures with Hugging Face, GPT, and NLP models for patient risk assessment.
Engineered scalable ETL pipelines across Amazon Redshift, Azure Cloud, and GCP to aggregate large healthcare datasets.
Applied customer segmentation models (KNIME, ML clustering) — increasing customer base by 5% and portfolio growth by 9%.
Cross-product adoption grew from 40% to 60% leveraging predictive insights and behavioral modeling.

Data Scientist

Mar 2024 – Feb 2025

Athenahealth · Milwaukee, WI

Improved claims classification accuracy by 20% through clustering and classification modeling using Scikit-learn and Spark MLlib.
Built ARIMA/ETS time-series forecasting models to analyze billing trends and improve financial forecasting for healthcare providers.
Implemented NLP techniques including Word2Vec and GPT-based summarization to enhance EHR search and automate clinical documentation.
Applied tree-based models with hyperparameter tuning via Random Search to optimize prediction and reduce denial rates.
Leveraged Spark MLlib for distributed ML processing, reducing model-training time and improving scalability.
Built Tableau dashboards integrating Oracle and SQL data to present operational KPIs and predictive insights to executive stakeholders.

Data Scientist

Jan 2022 – Nov 2023

American Express · Mumbai, India

Conducted large-scale customer behavior analysis using RFM segmentation and K-Means / Hierarchical Clustering to identify high-value and at-risk customers.
Designed personalized recommendation systems using Collaborative Filtering and Gradient Boosting Trees — improving engagement and acquisition.
Reduced default rates from 6% to 3% through advanced statistical modeling and optimized customer risk profiling.
Achieved recall rates as high as 94% via ROC-AUC evaluation, SMOTE oversampling, and cost-sensitive ensemble algorithms.
Designed and executed A/B testing frameworks to optimize marketing strategies and product offerings.
Cross-product adoption grew from 40% → 60%; customer satisfaction improved by 60% through advanced targeting.

Data Analyst

Jun 2019 – Dec 2021

DHL · Mumbai, India

Built PySpark ETL pipelines for data extraction and integration from SQL Server, reducing data latency and speeding up analyst access.
Designed demand forecasting models assessing sales demand, price elasticity, and operational risk for supply-chain planning.
Developed classification models (Logistic Regression, SVM, Random Forest, AdaBoost, Gradient Boosting) to optimize customer discount strategies across millions of records.
Applied SMOTE, PCA, and feature engineering to address class imbalance and improve model recall for logistics datasets.
Developed Tableau dashboards to visualize KPIs, demand forecasts, and campaign performance metrics for real-time stakeholder monitoring.

Murali Krishna
Bezawada

Numbers That Speak

Work Experience

Certifications

Skills & Tools

Education

Get In Touch

Murali KrishnaBezawada

Numbers That Speak

Work Experience

Certifications

Skills & Tools

Education

Get In Touch

Murali Krishna
Bezawada