Joydeep Swarnakar

Designed and deployed an end-to-end ELT pipeline transforming raw transactional e-commerce event records into high-performance Kimball Star-Schema Data Marts (dim_customers & fct_orders) backed by automated CI/CD schema assertions and an executive BI dashboard.

• Built and deployed NLP sentiment analysis models on customer feedback data spanning hundreds of thousands of records using NLTK, Hugging Face Transformers, and scikit-learn.
• Developed and deployed an NLP-powered chatbot to automate first-line customer support responses, reducing manual query handling and improving response consistency.
• Designed and maintained interactive Power BI and Tableau dashboards translating raw customer and operational data into business-ready insights used daily by 50–60 stakeholders across departments.
• Built and managed SQL-based ETL pipelines to extract, clean, and transform customer and operational data from multiple sources, ensuring reliable, analysis-ready datasets for downstream modeling and reporting.
• Developed forecasting models to support business planning using statistical and machine learning techniques on historical operational data.
• Applied computer vision techniques (OpenCV, PyTesseract) to automate extraction of information from image-based and scanned documents, reducing manual data entry effort.
• Ran A/B tests to evaluate the impact of model and feature changes, using results to guide iteration on deployed NLP and forecasting solutions.
• Administered and optimized relational databases (SQL) supporting core data science workflows, ensuring data integrity and query performance.

Supported digital marketing campaigns targeting student audiences, applying foundational digital marketing and audience segmentation skills in a live business setting.

Assisted in planning and executing outreach campaigns to support organizational fundraising and community awareness goals.