Srinivas M.

Data Scientist @Lowe's Companies, Inc.

US

Developed a Real-Time Stock Price Prediction model using LSTM neural networks in an Agile environment. Engaged with stakeholders for requirement gathering, reducing clarification time by 30%, and defining use cases for market predictions.• Utilized Python and TensorFlow/Keras for LSTM model development, improving prediction accuracy by 15%. Applied pandas and NumPy for data preprocessing, extracting features from historical data, and visualized trends with matplotlib for actionable insights.• Applied advanced SQL techniques (window functions, CTEs) to aggregate market data efficiently. Optimized queries, reducing data retrieval time by 25%, enhancing integration speed with financial data pipelines, ensuring seamless model data input.• Incorporated sentiment analysis using BERT for NLP, enhancing stock price predictions by 18%. Analyzed news and social media sentiment, combining insights with historical financial data to improve model performance and stock prediction accuracy.• Conducted hyperparameter tuning with GridSearchCV and RandomizedSearchCV to optimize the LSTM model\'s parameters. Improved performance and cut training time by 20%, raising prediction accuracy and model efficiency for real-time forecasting.• Performed testing using pytest and cross-validation (scikit-learn) to evaluate model accuracy. Conducted A/B testing, improving model robustness by 22%, ensuring consistent performance across stock tickers and various market conditions.• Deployed the final LSTM model using AWS S3 for data storage and AWS Lambda for serverless execution. Integrated AWS CloudWatch for real-time monitoring, achieving 98% uptime and triggering alerts for model anomalies.

University College of Engineering Narasaraopet - JNTUK

Bachelor of Technology - BTech

The University of Texas at Dallas

Master's degree