AI/ML Engineer - Risk Modeling & NLP

Durga Bhavani

AI/ML Engineer Irving, Texas +1-469-***-**** ***************@*****.*** LinkedIn

SUMMARY

AI Engineer with 5+ years of experience designing and deploying machine learning and deep learning models across insurance, banking and healthcare domains. Hands-on experience with TensorFlow, PyTorch and Keras alongside classical ML techniques (XGBoost, Random Forest, Logistic Regression) for risk modeling, anomaly detection and predictive analytics. Skilled in developing end-to-end ML pipelines using Python, SQL and cloud platforms (AWS, GCP) including AWS SageMaker for model training and deployment. Proficient in supervised and unsupervised learning and model evaluation techniques. Experienced in MLOps practices including MLflow, pipeline automation and model monitoring. Hands-on exposure to NLP and LLM-based solutions for automation and insights generation. Proven ability to deliver data-driven AI solutions that improve business outcomes and operational efficiency.

SKILLS

Programming: Python, SQL, Pandas, NumPy

ML & AI: Supervised & Unsupervised Learning, TensorFlow, PyTorch, Keras, XGBoost, Random Forest, Logistic Regression, Clustering, Anomaly Detection, Feature Engineering, Model Evaluation, Model Explainability (SHAP)

AI / NLP: Natural Language Processing (NLP), LLM fundamentals, prompt engineering, text classification, information extraction

Data Engineering & Pipelines: Data preprocessing, ETL/ELT pipelines, feature pipelines, ML-ready datasets

MLOps & Deployment: AWS SageMaker, MLflow, model monitoring, experiment tracking, pipeline automation, REST API integration, reproducible workflows, CI/CD basics

Cloud & Platforms: AWS (S3, Glue, Athena, Redshift), GCP (BigQuery), Snowflake

Statistics & Analytics: Hypothesis testing, A/B testing, time-series analysis, statistical modeling

PROFESSIONAL EXPERIENCE

AI/ML Engineer State Farm USA Jan 2024 – Present

Built XGBoost, Random Forest and TensorFlow-based deep learning models for risk scoring and claims anomaly detection, improving fraud detection precision by 22% and reducing false positives by 18%.

Designed ML pipelines using Python and SQL to process 1M+ claims records, reducing processing time by 30%.

Applied clustering and anomaly detection techniques to identify high-risk claim patterns, reducing manual review effort by 25% and accelerating case handling.

Optimized models through hyperparameter tuning and validation, improving model stability and performance by 20%.

Integrated ML outputs via APIs into business workflows, reducing decision turnaround time by 35%.

Conducted A/B testing and statistical analysis, increasing underwriting efficiency by 15% and improving model adoption.

Explored NLP/LLM-based solutions for document classification and insights extraction, improving automation efficiency by 20%.

AI/ML Engineer Capital One USA Jul 2022– Dec 2023

Developed predictive models (Logistic Regression, XGBoost) for forecasting and anomaly detection, improving accuracy by 20%.

Engineered scalable ML pipelines handling large datasets, reducing model development lifecycle time by 25%.

Applied machine learning to optimize workflows, increasing automation efficiency by 18% and reducing manual effort.

Improved model performance through tuning and monitoring, reducing model drift and prediction errors by 20%.

Deployed ML solutions into production systems supporting 50+ users, improving decision-making speed and consistency.

Leveraged MLflow for experiment tracking and version control, improving reproducibility and reducing rework by 15%.

AI/ML Engineer UHG USA Jan 2022– Jun 2022

Built machine learning models for patient risk prediction using clinical and claims data, improving prediction accuracy by 15%.

Built data preprocessing and feature pipelines, improving dataset quality and model performance by 20%.

Evaluated supervised learning models using validation techniques, improving reliability and consistency of predictions.

Collaborated with clinicians and engineers to validate model outputs, improving care outcomes and decision accuracy.

Data Scientist Wells Fargo India Aug 2019– Feb 2021

Analyzed large-scale financial datasets to support credit risk modeling and forecasting, improving prediction accuracy by 18%.

Built statistical and machine learning models, reducing forecast errors by 12% and improving planning accuracy.

Engineered features from raw data, improving model stability and reducing variance by 20%.

Evaluated model performance using cross-validation and financial KPIs, ensuring robustness and consistency.

Delivered actionable insights to stakeholders, supporting data-driven financial decisions and strategy.

EDUCATION & CERTIFICATION

Master’s in Computer Science, University of Illinois Springfield, Springfield, IL

AWS Certified Machine Learning – Specialty

Contact this candidate