Raushan Kumar

Experience

• 

  Amazon

  May 2017 - Oct 2020

  ❖ Failed Keyword Prioritization ModelExamine unsuccessful customer search terms on Amazon and utilize an ML algorithm to associate suitable choices with these keywords. Additionally, conduct an analysis to differentiate between irrelevant and relevant keywords. Discover a distinctive indicator for recognizing high-impact keywords and construct a model for such influential terms.❖ Product URL predictionDesigned and deployed a Logistic model to differentiate between legitimate product URLs and irrelevant ones. This model incorporated with Random Forest for feature selection techniques, serving as input for an LR model.❖ Product CategorizerDeveloped an ensemble ML model using OpenCV and TensorFlow to classify products. The model draws predictions from two sources i) Product attributes (Title, category, etc.) and ii) Product images. The model achieved >95% accuracy, categorizing >80% of product pages without any manual effort.❖ Brand Page URL ValidationBuild and deployed an NLP-based classification model which was fetching variables from itself with the help X-Path scrapping method and classifying between brand page and product page URLs. The accuracy and coverage of the model were >84%. Show less ❖ Dashboard AutomationConstructed an interactive dashboard for assessing keyword trends and volumes. I utilized Tableau, SQL, and Excel to compute monthly and quarterly business impact through the dashboard and deployed it on the server to enhance team efficiency.❖ Generalized ScraperBuilt an NLP-based ML model to select the best X-path configuration based on the historic X-path collection. It tries to fit the historic X-path for that competitor on all data points and tries to find the best X-path for each data point.❖ Best Model SelectorCreate a tree of the model to predict the best X-Path for feasible data points (Title, Model, etc.) to scrape data from competitors. It was an ML model which was classifying the best model among SVM, Logistic, etc., and suggested the best X-Path for each data point. Show less

  • Lead Data Analyst

    Jan 2019 - Oct 2020

  • Data Analyst

    May 2017 - Jan 2019
• 

  AB InBev

  Oct 2020 - now

  - Collaborated with developers to build best AI tool CatExpert.ai for assortment mix & Plano in the market.- Engineered Python-based libraries for clustering, Exploratory Data Analysis (EDA), and more, reducing effort time for Data Scientists by 2 - 4 days.- Engineered a computer vision-powered software capable of recognizing products placed on shelves.Possesses strong expertise in designing, developing, and deploying machine learning models for diverse applications.- Demonstrate a strong understanding of Probability, Statistics, and Operations Research/Optimization techniques at a foundational level.- Hands-on experience in writing reliable, maintainable, secure, and performance-optimized Python code.- Exhibit a robust understanding of Azure Cloud Platforms, specializing in service-oriented microservice architecture and design.- Developed and executed queries and analytical frameworks to generate replicable outputs for key stores attributes such as adherence, uplift, growth of business. - Applied advanced statistical methods (Cluster Analysis, predictive modelling) to identify differentiated customer segments, improving business targeting strategies for stores. - Designed and maintained SQL databases, ensuring data integrity through rigorous cleaning and transformation processes, producing high-quality, ready-to-use datasets for feeding into predictive models and statistical analysis. - Maintained comprehensive notebooks, re-usable code, libraries with detailed documentation, ensuring reproducibility of data processes and effective knowledge transfer across the team. - Effectively managed stakeholder expectations by providing timely responses, setting clear project goals, and communicating complex findings in an accessible manner. - Collaborated cross-functionally to enhance existing processes, identifying optimization opportunities and improving workflow efficiency in data analysis and reporting. Show less ❖ Assortment RecommendationClustering & SKU identification- Utilize consumption and demographic data to categorize stores effectively.- Leveraged machine learning models to classify various types of SKUs e.g. revenue driver, seasonal, etc.Product Recommendation- Derive the ML algorithm to provide the best set of Skus/ products to maximize their revenue based on the cluster, consumption data, etc.- Develop an ML algorithm to provide the optimal selection of SKUs/products that will maximize revenue, taking into account factors like clusters, consumption data, and more.Space allocation- Allocate shelf space based on the ML algorithm to ensure the optimal visibility of products.❖ Product Automation- Developed a library to decrease workload and enhance overall productivity.- Executed code migration to the Python platform for enhanced performance.- Introduce a product designed for seamless assortment recommendations, requiring minimal human involvement. Show less

  • Senior Data Scientist

    Oct 2023 - now

  • Data Scientist

    Sept 2021 - Oct 2023

  • Associate Data Scientist

    Oct 2020 - Sept 2021