❖ Fast-tracked to Associate Manager/ Senior Data Scientist owing to high performance.❖ Anomaly Detection on Welding Data➢ Led a 3-member team in performing Anomaly Detection on welding data with 300k+ rows. Developed a streamlined process involving clustering, classification, and adaptable anomaly detection for various machines. Successfully met business expectations by identifying and classifying critical anomalies.❖ Order Cancellation Forecasting➢ Led a team of 3 members to… Show more ❖ Fast-tracked to Associate Manager/ Senior Data Scientist owing to high performance.❖ Anomaly Detection on Welding Data➢ Led a 3-member team in performing Anomaly Detection on welding data with 300k+ rows. Developed a streamlined process involving clustering, classification, and adaptable anomaly detection for various machines. Successfully met business expectations by identifying and classifying critical anomalies.❖ Order Cancellation Forecasting➢ Led a team of 3 members to successfully conclude the Order Cancellation Forecasting project for a global manufacturing company. Developed multiple classification models with a target metric of f0.5 score for different quantity buckets. This project will result in annual labor cost savings of over $2mn+. Enabled plant managers to track scheduled and forecasted quantity for optimized labor and material requirements.❖ Market Segmentation➢ Spearheaded a 4-member team in a fast-track project to create marketing regions using a 2-step clustering approach. The goal was to support the sales team in optimizing product pricing. Successfully implemented the methodology to generate regions that provided valuable insights for more effective pricing strategies.❖ Influential People Mining➢ Led a team of 3 members in creating a classification and web scraping project with the goal of identifying influential users in the client's database for targeted marketing. The client had access to user data but lacked the means to leverage it effectively. Developed a scraping script to gather data from the internet and constructed a training set used to build a classification model for maximizing the identification of influential individuals. Show less

❖ Segmentation and Churn Analysis➢ Developed a customer clustering model for a distribution company, enabling targeted marketing andperformance analysis through an interactive dashboard. Segmented customers based on experience,engagement, and digital adoption. Created KPI's for the dashboard, empowering branch managers andcountry heads to make informed decisions and take corrective actions.➢ Collaborated with the company's founder to develop an in-depth Churn Analysis… Show more ❖ Segmentation and Churn Analysis➢ Developed a customer clustering model for a distribution company, enabling targeted marketing andperformance analysis through an interactive dashboard. Segmented customers based on experience,engagement, and digital adoption. Created KPI's for the dashboard, empowering branch managers andcountry heads to make informed decisions and take corrective actions.➢ Collaborated with the company's founder to develop an in-depth Churn Analysis, leveraging approaches like Order Day Gap, Inactivity Period, and customer gap patterns. Conducted extensive testing over months to identify insightful churn behavior. Overlaid this analysis with customer segments to redefine the client's churn definition, providing valuable insights for customer retention strategies.❖ Multivariate Forecasting➢ Developed a Champion Challenger implementation of a Multivariate Time Series Forecasting model for a major manufacturing company. Achieved a 10% MAPE across months for top SKU's and 99% accuracy in unit sales for the top 2 dfu's over 6 months. This model significantly optimized cost savings in planning, logistics, and operations. Incorporated diverse features such as construction indexes and stock market data as leading indicators.❖ Univariate Forecasting➢ Developed and managed a Time Series Forecasting model to predict cement sales for a leading global construction materials company. Oversaw forecasting and team coordination for dashboarding. Facilitated client interactions and on-time project delivery. Collaborated with another team for international implementation. Optimized code base and introduced multiprocessing for accelerated project completion. Deployed models on AWS, enabling enhanced production planning for the client. Show less

❖ Developed a predictive model to determine individuals' likelihood of traffic rule violations, aiding in insurance premium pricing decisions.❖ Leveraged Fuzzy Clustering to group customer data based on probabilities, identifying distinctive characteristics for targeted up-selling and cross-selling recommendations.❖ Implemented Python code enhancements, optimizing performance, to successfully de-duplicate a massive dataset of over 2.5 million records in a collaborative project.❖… Show more ❖ Developed a predictive model to determine individuals' likelihood of traffic rule violations, aiding in insurance premium pricing decisions.❖ Leveraged Fuzzy Clustering to group customer data based on probabilities, identifying distinctive characteristics for targeted up-selling and cross-selling recommendations.❖ Implemented Python code enhancements, optimizing performance, to successfully de-duplicate a massive dataset of over 2.5 million records in a collaborative project.❖ Optimized Deepspeech model for Indian Accent, delivering Speech-to-Text conversion on clean audio with satisfactory outcomes.❖ Developed and hosted an end-to-end Speech Data Mining project on AWS using Python and Flask. Enabled audio file intake, offering customizable outputs such as sentiment analysis, keywords, tone, summary, wordcloud, and diarization. Integrated multiple speech-to-text and text analytics cloud services. Show less

❖ Developed models to predict Claim Propensity and Customer Churn, leveraging SQL to extract a substantial dataset. Conducted Data Wrangling to enhance features and implemented diverse models like Linear Regression, Logistic Regression, SVM, Decision Trees, Random Forest, LightGBM, XgBoost, and AdaBoost.