Srishti Rath

Machine Learning Intern at Tech Mahindra

• Developed and optimized Sentiment Analysis model using Python libraries- Pandas, Keras, Matplotlib to classify customer feedback into positive, negative, or neutral categories, enhancing product satisfaction analysis • Engineered robust Python-based ETL pipelines, structuring data workflows to streamline extraction and transformation processes; achieved a 30% reduction in processing time while enhancing accuracy of customer feedback analysis • Demonstrated effectiveness of system by promptly identifying and addressing over 120+ service complaints, facilitating improved and efficient solutions to customer queries, indicating client appreciation

Student Research Assistant at Texas A&M University, College Station

Analyzed and transformed large data sets using SQL, Python, and Excel to extract insights, improve data quality, and support decision-making. Designed and developed Tableau dashboards to present trends and key metrics, enabling cross-functional reporting and clear communication of findings.

Software Developer at Ericsson

Collaborated with a team of 6 to develop Trigger Handler, a Java/Spring Boot application for real-time data streaming via Kafka, PostgreSQL, and Elasticsearch, monitored by Prometheus; deployed with Docker. Enhanced platform with IoT capabilities, achieving 85% efficiency gains and addressing complex testing issues by designing scalable test automation. Streamlined operations using JIRA, improving efficiency by 60% and delivering updates to teams in India and Sweden.

Developed and optimized a sentiment analysis model using Python (Pandas, Keras, Matplotlib) to classify customer feedback into positive, negative, or neutral categories, achieving 91% accuracy. Built Python-based ETL pipelines to streamline extraction and transformation, reducing processing time by 30% and improving analysis accuracy. Addressed 800+ service complaints to demonstrate system effectiveness and improve customer support.

Led Agile Scrum development as Scrum Master for a web application that automates course evaluation summarization using OpenAI API integration; replaced manual processes, increasing efficiency by 60% with SDLC, documentation, and testing.

Designed an encoder-decoder seq2seq model with attention, performing extensive hyperparameter tuning to improve translation accuracy, achieving 86% accuracy.