THARUN REDDY

Applied machine learning techniques such as regression, classification, clustering, and deep learning to develop robust predictive models.
• Developed end-to-end machine learning pipelines in PyTorch and TensorFlow/Keras, from data preprocessing and augmentation to model training and evaluation, ensuring reproducibility and scalability.
• Leveraged Databricks to develop and orchestrate end-to-end ETL pipelines, enabling large-scale data processing and real-time analytics across multiple data sources.
• Implemented MLflow on Databricks to track experiments, manage model versions, and streamline collaboration across data science teams.
• Optimized Spark jobs in Databricks for faster processing of structured and unstructured datasets, improving model training times and reducing compute costs.
• Optimized data processing pipelines using Apache Spark, enabling the processing of large-scale datasets in parallel and drastically reducing computation time for complex analytics tasks.
• Deployed machine learning models on Azure Machine Learning and automated deployment within Docker containers, streamlining the process from development to production.
• Automated model deployment and scaling in Kubernetes, ensuring models handled fluctuating traffic by dynamically allocating resources as needed.
• Integrated Azure data services such as Azure Data Lake and Azure SQL Database to manage large datasets and create scalable solutions for business intelligence and predictive analytics.
• Performed data cleaning, transformation, and preprocessing using Hadoop ecosystem tools and Python, ensuring high-quality datasets for analysis and ML tasks.
• Developed and executed ad-hoc SQL queries, collaborated with analysts, and created reusable Python codebases to improve accessibility, efficiency, and consistency.
• Automated Tableau reporting processes and implemented data security best practices using User Filters and Row-Level Security (RLS) for sensitive information protection.
• Implemented advanced statistical analyses using SciPy, such as hypothesis testing, regression analysis, and probability distributions, to support data-driven business decisions.
• Handled time series data manipulation in Pandas, including time-based aggregations and resampling, to improve forecasting accuracy and business planning.
• Managed automated testing frameworks within Jenkins to validate data models, pipelines, and analytics code before production deployment.
• Used Git/GitHub to track and manage the evolution of analytical models, ensuring version control and smooth collaboration across teams.
• Provided training to team members on R programming, data manipulation, visualization best practices, and model interpretation to improve overall data literacy.
• Explained model results to non-technical stakeholders by visualizing feature importance and interpreting complex ML models for better business understanding.

• Designed interactive and visually appealing data visualizations using R’s ggplot2 and Plotly, enabling stakeholders to quickly interpret insights and trends from large datasets.
  • Partnered with data engineers to architect and build scalable SQL-based ETL pipelines, ensuring smooth and reliable data flow from raw sources to structured formats for analysis.
  • Developed automated data extraction and transformation tools using Pandas, streamlining workflows and accelerating time-to-insight for complex datasets.
  • Built and deployed scalable machine learning pipelines on AWS Glue, automating ETL and data preparation processes to support model training and inference.
  • Applied machine learning algorithms using Spark’s MLLib for distributed datasets, improving the speed, scalability, and accuracy of predictive analytics.
  • Leveraged Databricks to streamline distributed data processing and ETL workflows, enabling faster and more efficient handling of large-scale datasets.
  • Implemented transfer learning in TensorFlow, leveraging pre-trained models such as ResNet to reduce training time and enhance accuracy for image and text classification tasks.
  • Used Keras to evaluate and validate deep learning models, ensuring robustness and generalizability before production deployment.
  • Developed custom loss functions and optimizers in TensorFlow to handle business-specific challenges, such as imbalanced classes and non-standard data distributions.
  • Maintained robust Jenkins pipelines to automate complex data workflows, ensuring seamless integration, testing, and continuous delivery of data-driven applications.
  • Deployed scalable data pipelines with AWS services such as S3, Glue, and Redshift, supporting efficient ETL processes and high-volume data analytics.
  • Improved performance of Power BI dashboards for large datasets through query optimization, reducing load times and enhancing user experience.
  • Created automated reporting solutions using Power BI, enabling business users to access real-time reports without manual intervention and improving departmental efficiency.
  • Optimized Spark jobs through partitioning, caching, and broadcasting, maximizing memory efficiency and performance when processing large-scale datasets.
  • Monitored PyTorch model training with TensorBoard and other visualization tools, tracking performance metrics and detecting issues early.
  • Tuned XGBoost hyperparameters using grid and random search, enhancing model accuracy and reducing overfitting.
  • Collaborated with cross-functional teams to define data requirements and design SQL-based reporting solutions, delivering actionable insights for multiple business units.
  • Utilized SciPy, NumPy, and Pandas to develop integrated data analysis pipelines, addressing complex statistical and mathematical challenges.
  • Deployed multiple data science tools within Docker containers, ensuring consistent and reproducible development and testing environments.
  • Managed Kubernetes namespaces and access controls, isolating machine learning projects and ensuring secure, efficient resource allocation.
  • Implemented MLflow tracking on Databricks to manage model versions, monitor experiments, and improve collaboration across the data science team.
  • Collaborated effectively using Git branches and pull requests, allowing simultaneous feature development and smooth integration into the main codebase.

• Implemented end-to-end data analysis and ETL pipelines using Python and SQL, automating extraction, transformation, and loading (ETL) of large datasets to support analytics and machine learning workflows.
  • Automated repetitive data cleaning, preprocessing, and transformation tasks using Python and R, improving accuracy, efficiency, and team productivity.
  • Built and deployed scalable machine learning pipelines using AWS Glue, Spark MLlib, TensorFlow, Keras, and PyTorch for classification, regression, clustering, NLP, and time-series forecasting.
  • Applied transfer learning with pre-trained models (e.g., ResNet) and developed custom loss functions and optimizers to address business-specific challenges, such as imbalanced classes or non-standard data distributions.
  • Tuned XGBoost hyperparameters using grid and random search, optimizing predictive accuracy and reducing overfitting.
  • Managed cloud-based data pipelines using Azure Data Factory, AWS S3, Glue, Redshift, and Azure SQL, enabling large-scale ETL processes and high-performance ML model training.
  • Optimized distributed data processing workflows using Apache Spark and Spark SQL with partitioning, caching, and broadcasting techniques to maximize performance and resource efficiency.
  • Developed interactive dashboards using Tableau and Power BI, providing stakeholders with real-time insights, applying row-level security, and training non-technical users to create custom reports.
  • Utilized Docker to standardize development and testing environments, and implemented Kubernetes namespaces and access controls to isolate ML projects and manage resources securely.
  • Monitored model training using TensorBoard and other visualization tools to track performance metrics and detect issues early, ensuring robust model deployment.
  • Maintained version control and collaboration using Git/GitHub with branches and pull requests, enabling parallel development and smooth integration across teams.
  • Built integrated data analysis pipelines using Pandas, NumPy, and SciPy, including geospatial analysis with clustering, distance calculations, and spatial relationship modeling.
  • Automated reporting and analytics workflows with Power BI and Tableau, reducing manual effort and improving accessibility for business stakeholders.
  • Leveraged Databricks to orchestrate distributed ETL and ML pipelines, improving scalability and handling of large-scale datasets efficiently.
  • Implemented MLflow on Databricks to track experiments, manage model versions, and enhance collaboration across the data science team.
  • Collaborated with cross-functional teams to define requirements, design SQL-based reporting solutions, and deliver actionable insights for strategic business decisions.