● Engineered a comprehensive RFM analysis pipeline, boosting targeted marketing effectiveness by 20% through improved customer segmentation. ● Built advanced forecasting models for footfall and sales by integrating historical data with external variables resulting in 10-12% improvement in prediction accuracy across products. Project: Customer Segmentation using RFM modelling & K-Means. ● Developed an RFM analysis framework that segmented over 2000 customers based on purchasing behavior, improving targeted marketing efforts. ● Identified high-value customers by analyzing Recency, Frequency, and Monetary metrics, leading to a 8% increase in targeted campaign effectiveness. ● Enhanced customer retention by 12% through tailored marketing strategies based on RFM scores. ● Automated data processing for RFM calculation and analysis, reducing processing time by 20%. ● Provided actionable insights that optimized promotions, resulting in a 15% increase in customer lifetime value. ● Explored frameworks like MLflow for model logging & tracking, validation and explainability. Project: Sales Analysis using XGBoost & time-series forecasting. ● Built and compared XGBoost and SARIMAX models for end-product sales forecasting. ● Incorporated lag values and rolling averages to capture historical sales trends and seasonality, resulting in improvement in model performance. ● Extracted date-based features, including month, day of the week, and holidays, enhancing forecast accuracy by 10%. ● Optimized model performance by implementing hyperparameter tuning & cross-validation leading to reduction in prediction error. Project: Sentiment Analysis with Deep Learning using BERT. ● Designed and implemented a BERT-based sentiment analysis pipeline to classify text data into positive, negative, and neutral sentiments. ● Preprocessed large volumes of text by cleaning, tokenizing, and converting inputs into BERT-compatible formats using Hugging Face Transformers. ● Fine-tuned a pretrained BERT model on a labeled dataset using PyTorch, achieving high accuracy and robust generalization on unseen data. ● Evaluated model performance with precision, recall, F1-score, and confusion matrix, achieving notable improvements over baseline classifiers.

● Optimized SQL code through indexing, resulting in a 18% increase in processing speed. ● Streamlined data cleaning and preprocessing procedures, reducing processing time by 15% and enhancing predictive model accuracy significantly. ● Enhanced data quality by addressing missing values, outliers, and skewed distributions, achieving a 40% improvement in data accuracy. ● Conducted feature extraction to improve dataset insights, leading to a 20% increase in model performance.

● Extracted and presented filtered results using SQL, enhancing decision-making efficiency for stakeholders. ● Documented access procedures to streamline user experience, resulting in a 20% reduction in ticket resolution time. ● Analyzed datasets to identify trends, providing actionable insights that improved project outcomes by 25%.