Harsh Y.

We conducted a classification task aimed at predicting a liability customer's likelihood to purchase personal loans, significantly contributing to the success of upcoming marketing campaigns. Our approach began with thorough Exploratory Data Analysis (EDA) to extract insights from historical campaign data, enabling data-driven decision-making for future campaigns. We assessed multiple machine learning models, ultimately achieving a well-balanced precision of 77.98% and recall of 70.83%. This… Show more

We conducted a classification task aimed at predicting a liability customer's likelihood to purchase personal loans, significantly contributing to the success of upcoming marketing campaigns. Our approach began with thorough Exploratory Data Analysis (EDA) to extract insights from historical campaign data, enabling data-driven decision-making for future campaigns. We assessed multiple machine learning models, ultimately achieving a well-balanced precision of 77.98% and recall of 70.83%. This robust predictive model effectively targets potential loan customers, enhancing the efficiency and effectiveness of marketing efforts. Show less

We employed advanced deep learning techniques to forecast potential customer churn within a credit card customer base. Our team developed and trained a dense sequential Artificial Neural Network (ANN) model, utilizing binary cross-entropy loss and the Adam optimizer to ensure efficient learning. Through a comprehensive evaluation of various optimizers and model variations, we achieved an impressive 86.30% accuracy during training and a solid 84.64% accuracy during testing, demonstrating a… Show more

We employed advanced deep learning techniques to forecast potential customer churn within a credit card customer base. Our team developed and trained a dense sequential Artificial Neural Network (ANN) model, utilizing binary cross-entropy loss and the Adam optimizer to ensure efficient learning. Through a comprehensive evaluation of various optimizers and model variations, we achieved an impressive 86.30% accuracy during training and a solid 84.64% accuracy during testing, demonstrating a strong predictive capability and providing valuable insights for customer retention strategies. Show less

Working with a prominent multinational consumer goods enterprise with annual revenues exceeding $10 billion, we revamped a scalable pricing and promotions solution for over 3,000 Health and Hygiene products. This initiative resulted in optimized operations and significant efficiency gains. By integrating PySpark and Databricks, we achieved an impressive 80% reduction in model runtime, cutting it down from approximately five days to just 12 hours. Additionally, we implemented advanced machine… Show more

Working with a prominent multinational consumer goods enterprise with annual revenues exceeding $10 billion, we revamped a scalable pricing and promotions solution for over 3,000 Health and Hygiene products. This initiative resulted in optimized operations and significant efficiency gains. By integrating PySpark and Databricks, we achieved an impressive 80% reduction in model runtime, cutting it down from approximately five days to just 12 hours. Additionally, we implemented advanced machine learning techniques that dramatically reduced the weighted Mean Absolute Percentage Error (MAPE) from 35% to under 20%, significantly enhancing demand forecasting accuracy. Show less

Working with a prestigious American multinational in the food and beverage sector with annual revenues exceeding $50 billion, we spearheaded transformative initiatives that optimized Go-To-Market (GTM) budgeting, resulting in a 9% ROI improvement across eight Latin American markets. Additionally, we developed a highly accurate Automated Invoice Tagging system and engineered a Share of Market prediction application. These innovations collectively saved $6.2 million USD through substantial cost… Show more

Working with a prestigious American multinational in the food and beverage sector with annual revenues exceeding $50 billion, we spearheaded transformative initiatives that optimized Go-To-Market (GTM) budgeting, resulting in a 9% ROI improvement across eight Latin American markets. Additionally, we developed a highly accurate Automated Invoice Tagging system and engineered a Share of Market prediction application. These innovations collectively saved $6.2 million USD through substantial cost reduction and risk mitigation. For further details on this transformative journey, please refer to the full case study here. ( Link : https://www.sigmoid.com/case-studies/franchise-transformation-with-data-products-optimizes-joint-budget-planning/ ) Show less