Developed an end-to-end baseline model to optimize procurement workflow and inventory management.• Designed and implemented a sophisticated data pipeline in Databricks to enhance data digestion in the model.• Conducted exploratory and statistical analysis of Xlarge-scale inventory, consumption, supply, and production data using Spark and Databricks on Microsoft Azure platform and developed mathematical cost function. • Built an innovative integer programming optimization model with PuLP to routinely suggest ingredients reorder scheduling.• Enhanced procurement lifecycle significantly by automating the reorder scheduling, resulting in a daily saving of 9 man-hour.

Machine Learning TA

Developed an end-to-end recommendation system (basket analysis) to boost sales and enhance marketing strategies.• Engineered a robust data pipeline with Python and SQL to extract, transform, and load. • Scrutinized extensive financial datasets statistically to extract association rules and correlation analysis. • Built a Bayesian model with Apriori unsupervised machine learning to identify critical metrics, and clustered products based on significance of support and lift values to provide actionable business insight.• Delivered data-driven recommendations to finance and marketing departments, advocating for product bundling, promotions, and redesigning stores’ layout to increase sales, with estimated potential annual sales increase ranging from 800,000 to 2,600,000 USD.

Machine learning TA

Provided in-depth data analysis and data science solutions to facilitate data-driven decisions.• Led customer segmentation utilizing NLTK NLP sentiment analysis and K-Means clustering to optimize marketing campaigns. Analyzed customer feedback to identify sentiment trends and grouped customers based on similarities. Results showed an 8% sales increase on a year-over-year basis.• Conducted churn analysis leveraging Random Forest algorithm to predict customer attrition. Utilized machine learning techniques to identify key predictors of churn and develop an accurate predictive model. Results informed targeted retention strategies, reducing customer churn by 11% on a year-over-year basis.• Employed time series decomposition techniques, specifically seasonal decomposition, utilizing Python and R. Analyzed smart meter data to identify peak network usage hours and days, leading to more efficient resource allocation strategies. The insights gained resulted in a reduction of emergency network upgrades from 12% to 5%, enhancing network stability and reliability.• Conducted A/B testing for website design and marketing campaigns. Analyzed key metrics to optimize user experience and drive conversions. Presented actionable insights for strategic decision-making.• Developed several informative PowerBI interactive dashboards and conducted day-to-day analysis using SQL to facilitate data-driven decision-making.