Aman Sharma

Utilized Regression Models to Drive 6% reduction in Dollar risk on GMI categories by developing and implementing diverse predictive strategies to optimize Price Promo relationship. Applied Factor Analysis Unsupervised ML Clustering Techniques to cluster diverse food Labels into New Themes. Exhibited exceptional data management proficiency by extracting and integrating vital information from Nielsen, Numerator, Label Insight, and Milly databases. Developed Word2vec and Bert embeddings on food labelled data and trending words.

Developed a worldwide project 'Part Recommendation Engine' which assist CCC Agents to Order parts, by design & analyse symptoms from the text and historical consumption. Expanded the PRE tool to support Chinese, Korean, and Japanese languages using Bert-base multilingual Network, resulting in a remarkable 25% reduction in error rate for recommending proper replacements of damaged parts. Leveraged Logistic Regression, Boosting, Ensemble, ANN, and Bert models to optimize and boost overall model efficiency by 12%.

Planning and Implemented YOLOv4 for tracking systems, resulting in a notable 10% accuracy improvement and an 8% reduction in processing time. Conducted A/B testing to address competitive client challenges, delivering effective recommendations for clearer platform roadmaps and selecting optimal CNN Architectures.