Nithin G Ambati

At Wells Fargo, I lead the development of large-scale machine learning and data engineering solutions across critical banking systems. My work spans from building PySpark- and Kafka-driven ETL pipelines to designing real-time streaming frameworks that support continuous analytics. I develop and deploy ML models on both AWS SageMaker and GCP Vertex AI, while ensuring every model is fully monitored, versioned, and governed using MLflow and internal MRM standards. I also build secure, containerized inference services on Kubernetes and automate CI/CD workflows through GitHub Actions and Jenkins. Alongside engineering work, I collaborate closely with risk, compliance, and analytics teams to ensure every solution is production-ready, reliable, and aligned with regulatory requirements

At Cardinal Health, I owned end-to-end ML pipeline development for large healthcare datasets, from ingestion and preprocessing to deployment and automated retraining. I designed Airflow- and SageMaker-based workflows, developed predictive and deep learning models for clinical and operational use cases, and supported near-real-time analytics through Kafka and Spark. I unified SQL, Snowflake, and AWS data sources to create high-quality, feature-rich datasets for modeling. My role also involved building MLOps foundations with MLflow, Docker, Kubernetes, and CI/CD tools, ensuring every model was reproducible, scalable, and monitored. I worked closely with clinical and analytical stakeholders to translate data into insights that improved patient outcomes and operational efficiency.At Cardinal Health, I owned end-to-end ML pipeline development for large healthcare datasets, from ingestion and preprocessing to deployment and automated retraining. I designed Airflow- and SageMaker-based workflows, developed predictive and deep learning models for clinical and operational use cases, and supported near-real-time analytics through Kafka and Spark. I unified SQL, Snowflake, and AWS data sources to create high-quality, feature-rich datasets for modeling. My role also involved building MLOps foundations with MLflow, Docker, Kubernetes, and CI/CD tools, ensuring every model was reproducible, scalable, and monitored. I worked closely with clinical and analytical stakeholders to translate data into insights that improved patient outcomes and operational efficiency.

At Citco, I developed machine learning models and data pipelines that supported major product and analytics initiatives. I worked extensively with Spark, PySpark, AWS Glue, and Athena to process large datasets, and automated training and scoring pipelines using Airflow. I deployed ML services with Docker and Kubernetes and standardized model development using MLflow for tracking and versioning. My work also included creating dashboards and visualizations to communicate model performance and business impact to engineering leadership. Throughout this role, I partnered with data engineers and software teams to integrate ML outputs directly into production systems.

At Novartis, I analyzed healthcare data to identify trends, improve processes, and support data-driven decision-making. I built ETL pipelines, automated reporting workflows, and performed statistical evaluations using Python and SQL. I created dashboards in Power BI and Tableau, delivered insights to clinical and business teams, and ensured high-quality, validated datasets for operational reporting. My role combined data engineering, analytics, and business support, helping teams improve efficiency and accuracy across multiple processes.At Novartis, I analyzed healthcare data to identify trends, improve processes, and support data-driven decision-making. I built ETL pipelines, automated reporting workflows, and performed statistical evaluations using Python and SQL. I created dashboards in Power BI and Tableau, delivered insights to clinical and business teams, and ensured high-quality, validated datasets for operational reporting. My role combined data engineering, analytics, and business support, helping teams improve efficiency and accuracy across multiple processes.