About Me
Experienced data scientist with a strong foundation in analytics and modeling, holding a Master's degree from Valparaiso University and a Bachelor's in Computer Science and Engineering from Hyderabad Institute of Technol…
Experienced data scientist with a strong foundation in analytics and modeling, holding a Master's degree from Valparaiso University and a Bachelor's in Computer Science and Engineering from Hyderabad Institute of Technology and Management. Hands-on experience as a Programmer Analyst at Cognizant, specializing in data analysis, model development, and deployment. Proficient in Python programming and libraries including Pandas, Numpy, and TensorFlow, with a focus on machine learning and natural language processing (NLP). Strong background in using relational databases such as MySQL and MS SQL Server for structured data storage and management. Skilled in the development and analysis of machine learning models such as CNN, RNN, LSTM, and traditional classifiers like Naive Bayes, K-Nearest Neighbors, Decision Trees, SVM, and Random Forest. Expertise in big data technologies like Apache Kafka and microservices, and cloud platforms including AWS EC2 and Docker for scalable model deployment. Strong background in data privacy and compliance, adhering to GDPR and CCPA standards. Experienced in data visualization and interpretation using Tableau, and adept at various data processing and analysis techniques including feature extraction and normalization. Certified Data Scientist with a Nanodegree from Udacity, showcasing advanced skills in the field.
Experience
Programmer Analyst
Collected 5TB of data with SSIS (SQL Server Integration Services) for high performance data integration and loading into MS SQL Server data warehouse for centralized storage and management.
Improved data quality by 78% using data preprocessing techniques in Pandas, Numpy, and Scikit-learn.
Optimized model training through feature normalization and scaling.
Designed an 8-layer deep neural network in TensorFlow, achieving 82% validation accuracy.
Conducted comparative performance analysis of CNN, RNN, and LSTM models against baselines.
Implemented dropout and L1/L2 regularization to mitigate overfitting.
Extracted 23 key features from raw data, focusing on credit and transaction metrics.
Enhanced models with non-traditional features like social media sentiments via NLP algorithms.
Improved credit default prediction accuracy to 85%, achieving a 0.89 AUC score.
Ensured GDPR and CCPA compliance for data privacy in feature engineering.
Utilized 5-fold stratified cross-validation for model testing.
Deployed models using Docker for scalable AWS EC2 infrastructure.
Enabled real-time data handling with Apache Kafka and microservices.
Developed Tableau dashboards for accessible model diagnostics.