Tushar Ahuja
Data Scientist with 5+ years of experience engineering scalable machine learning models, clinical NLP pipelines, and production-grade AI solutions across enterprise healthcare and insurance environments.
About
Data Scientist with 5+ years of experience engineering scalable machine learning models, clinical NLP pipelines, and production-grade AI solutions across enterprise healthcare and insurance environments. Expertise in transforming EHR and claims data into clinical insights using PySpark, Databricks, Snowflake, XGBoost, and LLM/RAG frameworks.
Proven track record of deploying end-to-end predictive models and automating MLOps workflows on AWS/Azure, improving risk stratification, reducing operational costs, and enabling clinical decision support through explainability, drift monitoring, and compliant data handling.
- 5 years' experience
- Joined July 2026
Industry experience
- Healthcare
- Financial Services
- Professional Services
Work experience
Architected patient risk stratification models using Python and XGBoost on 90M+ electronic health records and longitudinal claims data, improving high-risk patient identification accuracy by 29% and lowering downstream care management costs. Built clinical NLP extraction pipelines using Transformers and semantic embeddings to process unstructured physician notes, reducing manual chart review time by 58%. Developed an enterprise RAG clinical retrieval platform using LangChain, FAISS, and semantic search across 5M+ policy documents, reducing query latency by 42%. Established production MLOps governance with MLflow, SHAP interpretability, and automated drift monitoring, reducing false-positive prediction errors by 24%. Created distributed feature engineering pipelines in PySpark/Databricks/Snowflake for multi-terabyte datasets, reducing execution times by 35%. Ran A/B experiments and hypothesis testing for preventive care outreach, increasing member engagement by 18%. Partnered with opera
Engineered supervised ML models using Python, SQL, and LightGBM to automate insurance claim classification, improving accuracy by 31% and reducing manual review costs. Built distributed ETL pipelines with Apache Spark and SQL for semi-structured healthcare datasets, cutting data preparation times by 47%. Developed clinical NLP models with NER and text classification to parse unstructured medical records, boosting ingestion throughput by 52%. Built statistical regression frameworks to analyze healthcare utilization trends, reducing resource allocation expenses by 12%. Deployed containerized scoring pipelines via Docker on Azure ML, shortening deployment cycles by 36%. Built dashboards in Tableau/Power BI for 20+ executive stakeholders to monitor KPIs, accelerating decision-making by 30%. Implemented REST APIs for real-time inference over 50k+ daily transactions with response times under 85ms. Optimized database schemas and complex SQL in PostgreSQL warehouses (15k+ records) to reduce da
Education
Certifications and awards· 10
Hire Tushar Ahuja today
Post a job and invite Tushar, or contact them directly. No hidden commission.








