Poonam Mishra

Successfully uploaded data onto the AWS Neptune graph database, conducting comprehensive graph data exploration using the Gremlin query language. Leveraged Metaflow pipelines to generate and derive graph network features. Developed and evaluated candidate models, including Logistic Regression, Random Forest, and XGBoost, in Python, achieving a substantial improvement in fraud detection. Designed an efficient flow for customer insights and implemented name matching and address matching algorithms.

Developed a robust Logistic Model for B2B product targeting, resulting in a threefold increase in lead conversions and generating a $20 million revenue opportunity. Delivered comprehensive AP data analysis using Hive and SAS, automating data validation and generating actionable insights. Led data science projects including building Credit to Account (CTA) processes and developing propensity models for British Gas, utilizing Teradata and SAS Enterprise Miner.

Worked on service level optimization and dynamic inventory rebalancing using SAS. Utilized time series procedures in SAS to aggregate transactional data and applied k-means clustering. Validated various time series models, including ARIMA, and implemented forecast reconciliation methodologies.

Worked on Bayesian adaptive methods for dose-finding studies and validated results using R code. Performed technical validation of Exact tests and conducted fixed sample studies for different designs, endpoints, and hypotheses, including sensitivity analysis.