Machine Learning Data Science

Roshni Ranjita Bhowmik

+1-517-***-**** ********@***.*** linkedin.com/in/roshnibhowmik github.com/Roshni-Ranjita Education

Michigan State University East Lansing, MI

Master of Science in Data Science, Department of Statistics and Probability Aug. 2024 – May 2026 Indian Institute of Technology Kharagpur Kharagpur, WB Bachelor in Architecture, Department of Architecture and Regional Planning (8.2/10) Jul. 2016 – May 2021 Experience

Senior Analyst in Data Science at Tiger Analytics (Chennai, India) Jun. 2021 – Jul. 2024 Pricing Model (Automobile Industry) Singapore, Hong-Kong, Peru, Colombia

• Develop machine learning models in Databricks using Bayesian Mixed-Effects Modeling to calculate price elasticity

• Performed market basket analysis to determine relationships and cross elasticity among various product types

• Optimized pricing of these parts using Pymoo model with various market constants and obtained price elasticity

• Formulated a simulator to visualize the impact of alternate pricing using price elasticity and market constraints

• Adapted an alternative pricing method based on sales volume for countries with minimal price variation

• Improved the quarterly profit by 4 to 7% across both wholesale and retail markets in various countries Churn Prediction (Automobile Industry) Singapore, Hong-Kong, Chile

• Developed churn prediction model with 83% accuracy using Light GBM for customer retention in car services

• Analyzed and reported probable churn reasons using feature importance and SHAP values for each feature

• Use Chi-Square to analyze the relationship between categorical variables such as customer segments and churn.

• Retained 18% of customers from churning in a year and ensured they returned for regular scheduled car services Demand Forecasting (Automobile Industry) Singapore, Hong-Kong, Chile

• Lead a robust time series model using SARIMAX and Lasso Regression to predict next month’s vehicular sales

• Achieved a demand prediction accuracy with an RMSE of just 4percentage leading to precise resource allocation

• Extracted economic indicators such as the Consumer Price Index to integrate with sales data to uncover trends

• Developed a pipeline in Azure Data Factory to run monthly, generating predictions for the upcoming month Filler Machine Dashboard (FMCG) Thailand

• Solely developed a comprehensive dashboard using agile methodology to monitor filler machine performance

• Increased the production by 20% by analyzing the causes of downtime and developing prevention strategies

• Implemented metrics like total running time, quality loss, and PP waste to evaluate individual filler performance Machine Learning Intern at Everlytics(Bangalore, India) May 2019 – Jul. 2019 Developing an Application to predict machine downtime India

• Produced a full-stack web application to monitor machine performance and assess machine health in a plant

• Applied k-means clustering, an unsupervised machine learning method, to detect anomalies and overloading

• Predicted machine health by analyzing and visualizing data to identify deviations from regular patterns Project

Recommendation Model

• Developed a streamlit app for wine recommendation using a dataset of 100 different wines produced globally

• Analyzed wine features and ratings using interactive libraries like Altair and Bokeh for better visualizations

• Implemented a clustering algorithm to recommend wines based on user preferences using the nearest distance Sentimental Analysis

• Develop a model to classify user feedback as positive or negative on feedback data by leveraging NLP techniques

• Applied tokenization, stop words removal, Bag of Words, TF-IDF, and Naive Bayes technique to classify sentiment

• Achieved approximately 80% accuracy on test data using basic model, highlighting words with stronger sentiments Technical Skills

Programming Language: Python, R, SQL, Pyspark, C++/C, Linux Developer Tools: Azure, Databricks, VDI, Hadoop, MySQL, MongoDB, Tableau, Git Libraries: Pandas, NumPy, Matplotlib, Altair, Bokeh, Seaborn, Streamlit, Sklearn, Light GBM, Bambi, Pymo, Pulp, NLTK, spaCy, Hugging Face, TextBlob, TensorFlow, BERTopic, Dplyr, tidyr, ggplot, stats

Contact this candidate