Ramu Devarapalli

Data Scientist with 4+ years delivering production-grade machine learning solutions across healthcare and enterprise platforms, processing multi-terabyte datasets and driving analytics initiatives. Expertise designing predictive modeling systems, transformer-based NLP pipelines, and deep learning architectures using Python, PyTorch, TensorFlow, and Spark. enabling scalable, data-driven decision intelligence across healthcare and enterprise systems.

Data Scientist at Humana (2025-01 – Present)

• Developed predictive risk adjustment models using gradient boosting on claims and clinical datasets, improving member risk scoring accuracy by 28% while supporting value-based care optimization initiatives.
• Engineered distributed data processing workflows using PySpark and Snowflake, transforming 2TB+ healthcare records daily and improving feature availability latency by 37% for downstream modeling pipelines.
• Built NLP pipelines leveraging transformer architectures to extract medical entities from unstructured clinical notes, increasing extraction accuracy by 26% and enhancing care management decision support systems.
• Implemented MLflow-based experiment tracking and automated deployment pipelines, improving model governance, auditability, and reducing release cycle timelines by 43% across regulated healthcare environments.
• Designed fraud detection models using unsupervised learning techniques, identifying anomalous billing patterns with 24% higher precision and reducing potential claim leakage across insurance processing systems.
• Deployed scalable inference services using Docker and Kubernetes, supporting batch and real-time predictions with consistent performance while maintaining system reliability across cloud-hosted healthcare applications.

Data Scientist at Aspire Technolab (2020-06 – 2023-07)

• Designed recommendation engines using collaborative filtering and ranking algorithms, increasing user engagement metrics by 25% and improving personalization effectiveness across digital platform customer interactions.
• Engineered batch processing workflows using Hadoop and Spark frameworks, managing large-scale datasets exceeding 600GB, improving data transformation performance by 39% across analytics-driven business applications.
• Implemented classification models for churn prediction using ensemble techniques, improving retention targeting accuracy by 27% and enabling proactive customer outreach strategies across subscription-based services.
• Built time series forecasting pipelines for revenue prediction, enhancing forecast reliability by 33% and supporting data-driven planning decisions across finance and operational business units.
• Automated data validation and preprocessing frameworks using Python-based pipelines, improving data quality consistency by 44% and reducing downstream modeling errors across multiple analytics workflows.
• Developed interactive reporting solutions using Power BI and Plotly, enhancing stakeholder visibility into KPIs and improving reporting turnaround efficiency by 36% across executive dashboards.

Master of Science in Computer Engineering – California State University, Fullerton (2023-08 – 2025-05)

Bachelor of Science in Computer Science – Sathyabama Institute of Science and Technology (2017-05 – 2021-08)