Accurate Short Text Classification Using Bi-LSTM, Developed a sentiment analysis engine to classify large volumes of user-generated content using NLP and deep learning, Built and trained multiple machine learning models; Bi-LSTM achieved the highest accuracy among them, Focused on enhancing data preprocessing techniques and feature engineering to improve model performance.
Enabled scalable sentiment tracking to support real-time public opinion monitoring on social media platforms Sales Data Pipeline and Reporting Dashboard, Designed and implemented an end-to-end data pipeline using Apache Airflow, Spark, and AWS S3 to process over 10 million records daily, Integrated transactional data from multiple sources (CSV, API, and SQL databases) into a centralized data lake, Built an automated reporting dashboard using Power BI to provide real-time insights into sales trends, performance, and anomalies, Reduced manual reporting efforts by 70% and improved decision-making for sales and marketing teams IoT Sensor Data Ingestion and Analysis Platform, Engineered a scalable pipeline to collect and process real-time IoT sensor data using Apache Kafka and Spark Streaming, Stored processed data in a NoSQL database (MongoDB) for rapid access and analytics, Implemented monitoring and alerting with Grafana and Prometheus to track system health and sensor anomalies, Enabled predictive maintenance capabilities by analyzing trends and anomaly detection in temperature and vibration data