Pavan B Email & Phone Number

New York, Ny, Us

Plainsboro, New Jersey, United States

• Designed and implemented data pipelines using Azure and Hadoop to process large datasets, improving data processing efficiency by 35%.• Automated ETL workflows using AWS Glue and Apache Airflow, reducing manual intervention by 40% and streamlining data ingestion and transformation from multiple sources.• Developed and maintained scalable storage solutions with AWS S3 and Azure Blob Storage, reducing storage costs by 20% while ensuring high availability and data security.• Optimized distributed data processing systems using Hadoop MapReduce and Apache Spark, enhancing real-time processing and reducing latency for large-scale data applications.• Deployed machine learning models using AWS SageMaker and Azure Machine Learning, improving predictive accuracy and speeding up model deployment by 25%.• Implemented real-time streaming data pipelines using Apache Kafka and Spark Streaming, ensuring low-latency data delivery across teams for real-time analytics.• Collaborated with cross-functional teams to develop machine learning pipelines using TensorFlow and Pytorch, optimizing models for NLP and data analytics tasks.• Monitored and optimized performance of data architectures on AWS EC2 and Azure VM, reducing system downtime by 30% and improving data reliability.• Ensured data governance and security compliance using IAM roles, encryption methods, and compliance standards across AWS and Azure cloud infrastructures.• Built interactive dashboards using Tableau and Power BI for real-time business insights, supporting ad hoc queries and enhancing data visualization capabilities for stakeholders.• Evaluated and integrated new technologies such as Elasticsearch, Presto, and MySQL to enhance query performance and reduce response time for large datasets.

Vellore, Tamil Nadu, India

• Collaborated on designing Standard Operating Procedures (SOPs) for Exploratory Data Analysis (EDA), focusing on quantile and power transformations to reduce data noise by 20%.• Developed machine learning models using Python, TensorFlow, and Keras for healthcare applications, achieving 90% accuracy for Parkinson’s disease classification and 99% accuracy for diabetes prediction.• Implemented scalable ETL workflows on Azure Data Factory, automating data ingestion and transformation, reducing manual intervention by 40%.• Designed distributed data processing pipelines using Azure Databricks and Apache Hadoop, improving data processing efficiency by 35%.• Built and managed real-time data streaming systems using Apache Kafka and Spark Streaming, ensuring low-latency processing of anomaly detection tasks.• Optimized performance and scalability of machine learning models on Azure Virtual Machines, reducing system downtime by 30% and improving data reliability.• Utilized Azure Blob Storage for secure and scalable data storage solutions, reducing costs by 20% while ensuring high data availability.• Developed interactive dashboards using Power BI, providing real-time insights from anomaly detection models for data-driven decision-making.• Applied advanced visualization techniques using Matplotlib and Seaborn, creating visual representations of feature distributions and model performance metrics.• Integrated Docker for containerizing models, allowing seamless testing and deployment in Azure Kubernetes Service for scalable machine learning applications.• Maintained data security and compliance through Azure Active Directory (AAD) and encryption methods, ensuring data governance across all pipelines.• Utilized machine learning techniques such as Support Vector Machines (SVMs) and Artificial Neural Networks (ANNs) for anomaly detection and classification tasks.

Hyderabad, Telangana, India

• Developed and supervised data collection using Python and SQL, gathering over 500 biker feedback entries for operational improvements. Applied NLP techniques with SpaCy and Pandas for content summarization.• Deployed an NLP sentiment analysis system on Azure, using Keras for model training, to assess employee sentiment and reduce attrition risks.• Used K-Means clustering with Scikit-learn to group similar feedback, enabling targeted retention strategies. This approach resulted in a 10% increase in biker retention through improved, data-driven employee engagement efforts.

Master Of Science - Ms, Data Science, 3.6

Bachelor Of Technology - Btech, Electronics And Communications Engineering, 7.71

Activities and Societies: Volleyball, Cricket.- Undergraduate studies helped me develope intrests in Computing, finance and Data science..