Oct 2024 – Present · 1 yr 8 mos · Kolkata, West Bengal, India

• Collaborated with a U.S.-based client’s data team to train and enhance an AI model analyzing transaction patterns achieving an 8.3% boost in R² score and improved forecasting accuracy.
• Utilized advanced SQL and regex for cleaning, matching, and extracting credit card transaction data identifying false negatives and reducing false positives for higher accuracy.
• Leveraged Power BI and SQL to visualize key financial trends, driving insights that informed strategic planning.
• Applied pre-trained NLP models for sentiment analysis on in-store customer facial expressions, tracking emotions and enhancing engagement.
• Developed heat maps from video analytics to identify high-traffic zones and peak hours, optimizing store layout and staffing.
• Built a custom OCR pipeline using Py Tesseract to digitize handwritten medical documents, integrating NLP-based NER and regex workflows that reduced manual effort by 80%.
• Engineered an ANPR system using YOLOv8 for fast and accurate license plate detection.
• Designed Power BI dashboards to track email campaign performance (open rate, replies), leading to a 15% rise in engagement and 12% growth in conversions.
• Tech Stack: SQL | Python | Power BI | Amazon Redshift | AWS S3 | AWS Lake Formation | NLP | Regex | Py Tesseract | YOLOv8