▫ Collaborated with stakeholders, cross functional teams to gather requirements, optimize, maintain and monitor the pipelines. ▫ Led the price-promo team as a squad lead, managing 3 data scientists and overseeing project planning and development using Jira. ▫ Developed and Deployed RAG LLM application to extract product, various promotional information bundle, cashback, discount, etc. from unstructured retail data. This resulted in a 9% increase in accuracy and 12% wider product coverage, enabling precise pricing insights across promotional uplift, baseline sales, and price elasticity. ▫ Deployed multiple pricing insights & ML models (price elasticity, price promotion, price decomposition, competitor analysis) in GCP from research to production utilizing MLOPs practices. ▫ Demonstrated expertise in production-level Python coding - OOP, SOLID principles, TDD, adhering to software development best practices, and conducting successful code reviews to enhance code quality. ▫ Identified and

• Designed and deployed real-time, global food risk identification system for the UK Food Standards Agency leveraging LLMs to automate 90%+ of the previously manual process. ▫ Built web scraping modules to scrape food recalls, outbreaks, border rejection from various data sources including non-English articles using Python ▫ Built NLP pipelines as micro services to handles Language translation, food article detection, Opinion articles detection, duplicate information flagging, food/feed/FCM classification using finetuned BERT models. ▫ Extract Food Name, Country of Origin and Hazard, manufacturer and other key information from the articles using spaCy & sentence embedding techniques. ▫ Setup pipelines as micro services and integrated with CI/CD pipelines using GitHub actions
• Customer Segmentation & Customer Lifetime Value Prediction (Ecommerce) ▫ Built models to predict customer lifetime value and future purchases using ensemble model random forest. ▫ Customer segmentation based on

• Credit Risk Scorecard Model ▫ Developed acquisition & behavioral scorecard models for credit card & home loan portfolio using SAS ▫ Conducted score cut-off analysis, contributing to 7% reduction in credit risk. ▫ Documentation of model development and validations of scorecards ▫ Collaborated with stakeholders to regularly review and revise the credit policy based on customer segmentation and risk profiling.

• Warranty Fraud Detection using ML which has a business impact of $20M. ▫ Segmentation of warranty claims into noncompliant, highly suspicious, suspicious and compliant claims employing advanced clustering algorithms such as fuzzy-c and k-means ▫ Developed predictive models to detect/predict noncompliant warranty claims by using machine learning techniques such as SVM, random forest, and naïve bayes classifier using
● • Warranty Units & Customer Service Order Forecasting ▫ Developed Forecast model to forecast customer service orders and warranty units using time series techniques and built a simulator tool in excel.
• Customer Experience Key Driver Analysis ▫ Built Structural Equation model (SEM) to estimate cause-and-effect relationships of satisfaction-loyalty chain and identified what is important to customer Satisfaction and Loyalty intentions. ▫ Developed SAS Macros for Shapley Value Regression, to measure the importance of attributes in key driver analysis.

Estimated time varying volatility for NSE stock market index using GARCH models. Investigated index return characteristics such as non-normality, fat tails, volatility clustering. Estimated volatility using different GARCH models and identified best model according to log likelihood value and to the diagnostic tests.
- Developed SAS Macros to extract data from WRDS database.
- Modeled Capital Asset Pricing Model to estimate the expected returns and cumulative abnormal returns for 3 and 5 days event window.
- Modeled market reaction and operating performance with deal factor using Regression.
- Computed Bid-Ask spread for NSE S&P CNX 500 companies.
- Computed Impact cost on a trade for Rs. 2, 5 and 10 million of the S&P CNX Nifty.
- Created minute-by-minute snapshot and second-by-second bestquotes.
- Created the Network for Security Markets Database (NSMD) for Spot and Derivative data for the period April 1999 - to Apr 2007, sponsored by the National Institute for Security Markets (NIS

Developed SAS macro to automate data extracting, importing and preparing SAS data in desired format for daily sales data.
- Cleaning, transforming, and reporting discrepancies in the data and resolved data quality issues.
- Computed inventories, summary statistics and depicting graphs and investigated trends and seasonality in sales data.
Estimated and forecasted demand using time series models.

Worked on project titled “Comparing tooth cleaning efficiency of brushes” with Prof. Asis Kumar Chattophadyaya (ISI) using Multivariate Statistical techniques MANOVA and CLUSTER analysis.
•Worked on project project titled “Principle Component Analysis using MM- Estimator and Robust Bootstrap” with Prof. Arni Srinivasa Rao (ISI).• Worked on project titled “Comparing tooth cleaning efficiency of brushes” with Prof. Asis Kumar Chattophadyaya (ISI) using Multivariate Statistical techniques MANOVA and CLUSTER analysis. •Worked on project project titled “Principle Component Analysis using MM- Estimator and Robust Bootstrap” with Prof. Arni Srinivasa Rao (ISI).