Developed and deployed Predictive ML Models (Linear Regression, Random Forest, Decision Trees, XGBoost, SVM, Neural Networks and Large Language Models (LLM)), enhancing product yield forecasting and manufacturing process efficiency by 30%.• Built a custom Named Entity Recognition (NER) classification model using TensorFlow for medical-grade health data, increasing accuracy from 89% to 95%, supporting critical healthcare data applications.• Integrated Retrieval – Augmented Generation (RAG) techniques using tools such as Haystack and LangChian to improve the precision and recall of text -based data extraction.• Designed and automated ETL pipelines in Dataiku for seamless data migration from Snowflake to Google Cloud Platform (GCP), implementing data validation checks, leading to 45% reduction in data processing time.• Deployed machine learning models and data workflows on GCP services (BigQuery, Dataflow) to enable real-time data analysis and scalable storage solutions, ensuring efficient handling of streaming data from wearable and IoT devices.• Created 15 Power BI dashboards for real-time tracking of manufacturing performance and developed a PowerApps application to improve communication of key performance indicators (KPIs) with internal stakeholders.• Utilized SQL with over 70 CTE queries in Genesys, optimizing analytical reporting and retrieval efficiency by 20%• Automated data wrangling, data modeling, feature selection, and model performance reports in Dataiku.• Worked closely with the Data Science team at Feinstein Institutes for Medical Research, contributing to a $3 million grantproject focused on high-volume, medical-grade data processing for real-time clinical monitoring.

• Built ensemble machine learning models including (Logistic Regression, SVM, Decision Trees, Random Forest, XGBoost, and GBM) using scikit-learn for churn prediction, reducing false positives & negatives by 27%.• Enhanced loan default prediction accuracy by optimizing Random Forest model, improving F1 score from 0.71 to 0.84 through targeted feature engineering and hyperparameter tuning.• Implemented XGBoost model using R programming with over 500 features for fraud detection, employing regularization and cross-validation to facilitate model selection.• Designed a SARIMA time series forecasting model in Python (pmdarima), achieving 83% MAPE in anomaly detection and impact analysis, and deployed on AWS SageMaker.• Automated monthly financial reporting processes using Python, SQL, and Tableau and standardized metrics & KPIs across departments, resulting in a ~72-hour monthly reduction in manual workload.• Implemented ETL pipelines for real-time data streaming of customer engagement data to AWS Kinesis Data Firehose and AWS S3 buckets and utilized AWS Glue and Athena for data transformation and extraction.• Built Python-based topic and sentiment analysis using NLTK library and Hugging Face's pre-trained BERT model (Large Language Model), providing dashboards for complex customer interaction data analysis.• Led data mining from AWS Redshift and Snowflake data lake using SQL and Python for data modeling and ad-hoc analysis driving insights for strategic decision-making.• Conducted data cleaning, preprocessing (one-hot encoding & target encoding), and Exploratory Data Analysis on ~20 million customer records using NumPy, Pandas, SciPy, Matplotlib & R Shiny for actionable insights.

• Developed a collaborative filtering recommendation system for an e-commerce platform, resulting in an 11% increase in conversion rates through rigorous statistical analysis (hypothesis testing) and A/B testing.• Collaborated with cross-functional teams to integrate Python, R, and SQL in traditional statistical modeling (ANOVA) for product feature analysis, enhancing user engagement and conversion rates on the website.• Led Bayesian A/B tests (hypothesis testing) for website design & recommendation algorithms, leading to a 9% improvement in click-through rates.• Executed ad-hoc data analysis & processing with Python, and Spark & utilized SQL queries for GCP BigQuery to extract sales trends & insights.• Defined success metrics for product analytics, conducted power analyses &provided actionable recommendations.• Applied Random Forest in Marketing Mix Modeling to analyze marketing expenditure, optimizing budget allocation for increased ROI.• Designed and published interactive data visualizations, dashboards & reports in Tableau, providing stakeholders with real-time sales and product insights for a retail client.• Produced predictive analytics reports using Python and Tableau, visualizing prediction outcomes and key performance indicators to inform strategic business decisions.