Surendhar Chowdary Email and Phone Number

• Developed a real-time financial analytics system using Python, Pandas, and NumPy, providing personalized investment recommendations based on user profiles and market data.• Designed and maintained ETL pipelines with Python and AWS Glue, ensuring efficient data integration from market feeds, user profiles, and financial news for real-time analysis.• Utilized AWS Lambda and Apache Kafka for effective real-time data ingestion and processing, enabling dynamic adaptation to incoming data streams and timely insights generation.• Implemented an ensemble machine learning approach using Gradient Boosting and Random Forest models in Scikit-learn for risk assessment and portfolio optimization.• Enhanced investment decision-making by applying advanced NLP techniques with NL TK and Gensim for sentiment analysis on financial news and social media, extracting actionable insights..• Managed model deployment on AWS SageMaker and orchestrated container management with Kubernetes, optimizing resource allocation and system responsiveness for large-scale data inputs.

Spearheaded the enhancement of the fraud detection system, leading 30% drop in false positives and 20% gain in detection accuracy. Implemented advanced ML models to improve fraud detection efficiency and reliability.•Performed EDA on a dataset of over 1TB, including transaction and user behavior metrics. Identified key patterns and anomalies using statistical techniques, leading 15% gain in model performance through improved feature engineering.•Processed and cleaned datasets exceeding 500GB using Python, SQL, and NoSQL. Achieved a 98% accuracy rate in data integrity through advanced imputation, normalization, and feature engineering techniques.•Developed and optimized machine learning models including Logistic Regression, Decision Trees, Random Forest, and Gradient Boosting. Utilized PySpark to train models on datasets of up to 1TB, achieving a model precision of 92% and a recall of 89%. Reduced model training time by 35% through efficient algorithm tuning and hyperparameter optimization.•Conducted model evaluation using Precision, Recall, F1-Score, and ROC-AUC. Enhanced fraud detection performance with an F1-Score increase of 10 points and ROC-AUC improvement of 15% through iterative model refinement.•Implemented CI/CD practices using Docker and Git, automating the model deployment pipeline. Reduced deployment errors by 40% and cut deployment time by 50%, ensuring more reliable and timely updates to the fraud detection system.•Created interactive dashboards and reports using Tableau and Power BI. Developed 10+ dynamic visualizations that highlighted key trends and model performance metrics, enhancing stakeholder decision-making speed by 25%.•Documented the complete model development process, including data preparation, model selection, evaluation, and deployment. Prepared detailed reports and presentations, improving stakeholder grasp and engagement by 30%.