As a Machine Learning Engineer at Rebecca Everlene Trust, I contributed to the Eat Like Us project, an initiative focused on understanding and reducing food waste in communities. My work involved collecting and preprocessing data from multiple open-source platforms, including public food-distribution datasets, consumer-behavior reports, and environmental data sources.

I supported the team in analyzing patterns of food usage and identifying regions and demographic groups where food wastage was most prominent. I also assisted in building early-stage predictive models to estimate surplus generation and potential redistribution needs. This helped guide the project’s goal of improving food-waste awareness and recommending more efficient consumption strategies.

Overall, my role combined data collection, feature exploration, basic ML experimentation, and insights generation to support a social-impact-driven machine learning initiative.

Spearheaded the development of 5+ ETL pipelines using Python and SQL to clean and transform 70,000+ sales and service records for client TVS Motors, resulting in a 35% reduction in manual processing time and improved reporting efficiency.

Developed and deployed 8 interactive Power BI dashboards to visualize SLA breaches, revenue trends, and performance KPIs, empowering strategic decisions by business leads

Performed exploratory data analysis (EDA) and prepared clean, structured datasets through feature engineering, encoding, and outlier handling, enabling accurate forecasting and anomaly detection models using Scikit-learn

Generated insights that improved reporting accuracy by 20% and directly influenced prioritization of underperforming regions

Engineered and maintained robust data validation frameworks using SQL to ensure integrity across SAP ERP , Manhattan WMS, and TMS platforms, overseeing the accuracy and consistency of more than 200,000 inventory records per week for a client company

Built internal reporting tools using a combination of Excel automation and custom SQL queries, enabling operations teams to monitor order volumes, fulfillment status, and stock reconciliation metrics across 30+ warehouses with improved response time

Identified and resolved recurring inventory mismatches and data synchronization failures through structured root cause analysis (RCA), leading to a 22% reduction in order processing errors and ensuring consistent upstream-downstream system alignment

Documented complex integration flows, including EDI interfaces and batch jobs, while collaborating closely with QA, DevOps, and deployment teams during release cycles to maintain system stability, data consistency, and audit compliance