Dhirendra Panwar

Data Scientist familiar with gathering, cleaning and organizing data for use by technical and non-technical personnel. Advanced understanding of structured data modelling, touched upon CV and NLP for few inbetween projects. Highly organized, motivated and diligent with significant background in banking, retail and ecommerce domain.

Tiger Analytics: 2020 - 2024

Propensity Modeling for Banking Client

• Developed propensity model for Marketing team to target primary account holders.
• Led problem formulation, data analysis, modeling, testing, and deployment.

Lead Qualification Algorithm Development

• Engineered algorithm for US Technology firm to prioritize high-conversion leads.
• Utilized tree boosting algorithms and used list and gain charts as final evaluation metrics.

E-commerce Product Subcategory Identification

• Identified profitable subcategories, boosting annual sales by $1M.
• Conducted exploratory analysis, utilized XGBoost, and created actionable insights.

Inventory Analysis for US Store Chain

• Scraped and analyzed product quantity data, achieving 88% sales prediction accuracy.
• Employed Median Absolute Deviation / Z-score methods for competitive benchmarking.

Sales Forecasting Model for Banking Client

• Developed model predicting nationwide sales for upcoming quarter.
• Implemented fbprophet library; assessed accuracy using MAPE metrics.

Virtusa: 2018-2020

Data extrapolation:

• Developed and maintained backward data filling pipelines for a banking client. Utilized proxy, cohort and EM algorithm for the same.

NLP email segmentation

• Developed an email segmentation model for auto classifying emails in the respective buckets based on their content.