Purchase Structure:- Developed a tree-like representation of the market predicting product competition for 6 countries &8 categories usingHierarchicalclustering&Chi-square tests, with share & growth rate calculated for each node◦ Provided strategic insights by analyzing consumer loyalty, market competition, pricing benchmarks, & optimizing assortments & brand positioning

- Managing a forecasting project for a major Beverage client, implementing hyperparameter tuning alongside 8 different time series algorithms to predict sales across 80 countries, 9 categories, diverse channels, brands, & trademarks. Improved accuracy by selecting the best model and automating processes, resulting in a 50% reduction in production time and a significant increase in efficiency- Completed a 5-day advanced regression training at IIM Bangalore, enhancing skills in effective regression analysis

- Developed a Streamlit web application utilizing the YouTube API to scrape and analyze data, emphasizing time‐related insights, sentiment analysis, topic modelling, summary generation, and engagement distribution over periods for product understanding. Implemented NLP, NMF, & LLM models to facilitate comprehensive analysis- Led a 3‐member team in scraping data from 100s of pages across 6 gaming sites using GCP OCR API, Beautiful Soup, Selenium, and Scrapy. Created concise dashboards with 10+ charts for Weekly, Monthly, Seasonal, Sentiment & Battle Pass insights using Google Sheets

- Developed different machine-learning models like Time series forecasting, Customer Segmentation and Sentiment Analysis using various machine-learning algorithms & helped businesses in the gaming industry- Visualization using tools like Excel, Looker & PowerBI helps businesses gain insights - Hands-on experience with popular algorithms in supervised and unsupervised learning- Cloud based(AWS, Microsoft Azure and Heroku) deployments of ready-to-use models- Experienced in handling 50 M records- Worked on scraping projects from dynamic and static websites using different libraries like selenium, beautiful soup, scrapy etc- Player Segmentation >The project involved analyzing player-churn reasons using segmentation; 40 KPIs were selected and used to build an analytical dataset. >A player segmentation model was then developed using the k-means clustering algorithm and factor analysis. The goal of the model was to segment the players based on their playing behaviour and provide valuable insights to the client and achieved a good silhouette score.>The final solution was delivered to the client in the form of a deck, presenting the findings and recommendations.- Time Series Forecasting> The project aims to forecast player retention and Average Revenue per Active User (ARPWAU) for 720 days. Time series forecasting techniques were used and Exploratory Data Analysis (EDA) was performed to examine the data and identify patterns.> The stationarity, trend, seasonality and outliers of the data were checked and handled, Several time series models were tried and tested, and FB Prophet was chosen based on the MAPE

> Insurance premium prediction- Objective: The goal of the project was to build a machine learning model to predict insurance premiums.- End-to-end regression model: A comprehensive regression model was developed to make predictions based on various factors related to insurance.- CI/CD pipeline: To streamline the model building and deployment process, a Continuous Integration and Continuous Deployment (CI/CD) pipeline was set up using Github Actions and AWS.- Storage: The model and its dependencies were stored in an S3 bucket and an ECR repository for easy access and management.- Webapp creation: To make the predictions of the model accessible to users, a web application was created using the Streamlit framework.- Deployment: The web application was deployed, making it available to users.- Automation: The entire process was automated, from model building to deployment, making it more efficient and streamlined.

• Managed & reported on an Oracle database. Suggested automation for recurring jobs actively supported during deployments• Conducted log analysis to troubleshoot errors, established root causes of application errors and escalated serious concerns to the Senior Engineer