Srinivasa Karur

• Developed a dynamic pricing model using Random Forest & Gradient Boosting, increasing revenue by 15% in 3 months.
• Automated pricing pipeline using Flask REST API; reduced manual interventions by 80%.
• Built CI/CD for deployment on AWS; implemented real-time monitoring via Power BI.
• Conducted A/B testing, customer feedback analysis, and retrained model for continuous improvement.
• Role included gathering business requirements, architecture, proposing techniques, frameworks and algorithms, and Data collection,data ETL process,Model development ,Deployment , visualization and monitoring.

• Forecasted monthly product sales using LSTM, SARIMA, and GBM models, improving inventory accuracy.
• Conducted EDA, feature engineering, and seasonal trend analysis across 5 years of sales data.
• Deployed forecasting API via Docker on AWS EC2; integrated monitoring for MAE and MAPE.
• created dashboards in Power BI; deployed models on AWS EC2 via Docker

• Built Generative AI solution with fine-tuned BERT to highlight key document sections.
• Deployed search interface using Flask and Docker Compose; reduced search time by >60%.

• Designed effective prompts for LLMs (OpenAI) for sentiment analysis, summarization, multilingual response.
• Created prompt library and evaluation pipeline (BLEU, ROUGE, Human Eval) for production readiness.
• Deployed web interface and containerized solution on AWS cloud.

• Built foundational data pipelines using Kafka, Spark, and Flume for B2B customer analytics.
• Enabled real-time ingestion and transformed data for marketing & engagement dashboards.

• Ensuring high-quality data for targeted marketing campaigns and comprehensive data analytics.
• Delivered real-time campaign analytics using Airflow, Snowflake, Spark, Power BI.
• Built end-to-end pipelines from e-commerce and social data sources; drove 18% campaign efficiency.
• projects aimed at enhancing business revenue,customer success, and operational excellence including Reports on customer acquisition, engagement, and retention

• Customer Segmentation for Bank Marketing Strategies: The goal of this project is to study and predict the right group of new customers for a Bank, so the company can adopt the specific proven marketing strategy to each of them and be more successful in the business. Based on all classifiers score.
• Developed customer segmentation using AdaBoost, Decision Trees; improved targeting by 25%.
• Delivered actionable insights via Power BI and Jupyter-based visual reports.

• Built recommendation engine using hybrid collaborative filtering and popularity-based model.
• Boosted conversion by 20% and reduced RMSE by 15% through tuning and A/B testing. Boosted user engagement metrics by 25%,

• Built electricity price prediction model with time-series data using XGBoost and LSTM.
• Created ETL pipelines and deployed model via MLflow; visualized trends via Prometheus + Grafana.

• Created Flask-based ML app for healthcare cost prediction, deployed on Azure using Docker.
• Built regression pipelines, conducted EDA, integrated CI/CD and API services.

Implemented project Business Planning and forecasting by using SAP BPC10.1 on BW7.40(Datawarehouse) on HANA. The project consists of implementing Datawarehouse and upgrading SEM /BW to BW 7.40 on HANA(Cloud)

Implemented SAP Business warehouse for Inventory Optimization of MM, IM, Fixed Assets, Capital Spares from Plant and property, Plant &Equipment, ASSETS or Fuels for Working capital, IM, Purchasing.

The EMM -Enterprise Marketing Management project implemented to provide Real -time spending tactics/commitments

Implemented to Split the systems as per two separate legal entities to support Reporting System (MDLZ and Kraft) for about 3 thousand users across US and Canada utilizing BI/BW/ HANA/BOBJI /Portal Platform.

To provide the business the required BO universes and required reports for Customer service and Maintenance Order Notification, Operation and Header with info from cost, User status, Maintenance plan, Customer service Order and PM Order Header info and also required Master data and its attributes.

The project Sales and Trade Reporting System implemented and RTR (Record to Report) system for about five thousand users across US and Canada utilizing BI/BW, Business Object Platform.

The project consists of 5 releases with first release consists of financial and HR reporting, second release consists of Business consolidation and HR Reporting, third release consists of Business planning reports, fourth release consists of CRM-Analytics - CRM-Marketing reports and fifth release consists of CRM-Analytics CRM-Sales and CRM-IPM Reports.

Reports on Monthly reports on Demand Plan, Budget, and Forecast data and drill down to daily level and reporting including web reports and graphs, reconciliation reports

The scope of the Project is to develop “Management Information system” for Utility Industry with Development of Business ware house objects and “Management Cockpit and Balance scorecard “

The project implemented with scope to develop “Executive reporting System” by automation of data collection from both SAP R/3 & non-SAP systems and other multiple ERP systems

The project implemented with scope to integrate Direct and Indirect sales data where Direct sales data is from source SAP and indirect sales data from non-sap source, and modification to the existing design and new integrated reports.

The project implemented with scope to integrate SAP and Non-SAP Balance sheet and Profit and Loss data and to develop individual and integrated balance sheet and profit and loss reports for both sap and non-sap companies

• Reviewed - Design of Back End objects, Architecture, Query performance and to implement major changes to back-end objects with Logical partitioning based on segments