Avirup Sen Gupta Email & Phone Number

Plantation, Florida, Us

• Created an in-house simulation engine for finding optimal number of SKU in Chewy product catalog.• Trained an ML model for Inventory forecasting at various Merch Class levels at one through twenty-five weeks out.• Acted as a sounding board and mentor for more junior members of the team – provided guidance and best practices to the group which has yielded better outcomes from members of the team. • Led the Inventory Analytics Team for Chewy’s safety stock models development.• Built reporting tools for monitoring Safety Stock model Performance, which led to an improved understanding of dynamic relationship between demand uncertainty, fill rate and service level.• Developed a predictive model for estimating purchase order arrival probability and timing.• Developed a stochastic simulation tool for scenario analysis.• Perform A/B test for optimizing safety stock refresh cadence and demand variability formulation.• Develop simulation-based model for finding optimal Autoship (subscription) reserve days.

Santa Monica, Ca, Us

• Played primary owner role within the company for the data science solutions to support client’s pricing, underwriting and risk selection models as well as strategic data needs for future improvements. • Apply Generalized Linear Models (Poisson, Gamma and Tweedie) in order to calculate the auto insurance pure premium given the observable characteristics of the policyholders.• Applied multiple machine-learning classification algorithms to categorize the users (as an underwriting tool) for a worker compensation cell phone app used at various construction sites.• Forecasted flood for small to medium sized watersheds applying a Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) neural network model on daily discharge, rainfall, temperature and other meteorological data as inputs.• Applying clustering methods to classify territory for liability and collision insurance ratemaking• Analyze the impact of telemedicine (such as frequency, severity, reporting lag, closure rates, attorney involvement, job transfer rate and return to work) in workers’ compensation.

New York, Ny, Us

• Query structured/semi-structured data and perform exploratory data analysis on customer data and provide actionable insights.• Classify high-risk accounts/customers from the medium-to-low risk customers using supervised classification algorithms.• Work with large scale unstructured exposure and catastrophe hazard data and fuse them with other enterprise data sources (such as insurance claims) to quantify, and/or forecast business metrics (such as pure premium).• Lead the development of fully automated flood response systems that can estimate event’s impact on insured property, insurers business and the supply chains they insured as well as the overall financial impact of catastrophe over various business units.• Developed a geospatial analytics engine in R for identifying impacted exposures by overlaying real-time catastrophe event "footprint" on the entire portfolio of insured exposures. • Use Artificial Neural Network (ANN) to forecast flooding in Odra and Vistula River Basins and validated the accuracy of the model against actual flow.• Using various optimization techniques (such as Gradient Descent and Particle Swarm Optimization) for hydrological model calibration.

Boston, Massachusetts, Us

• Performed statistical and hydrological modeling to estimate spatial-temporal scale, frequency, severity of flood events.• Developed visualization materials for making inferences from large data and gaining business insights using R, MapWindow and ArcMap desktop applications.• Automated spatial and predictive analytic models using R script and ArcGIS model builder tool.• Collected, processed and integrated highest quality spatial and time series data (e.g., elevation, hydro-climatic, soil, land use and land cover) into company’s modeling framework from government and proprietary data sources.• Developed and implemented statistical methodologies for validating precipitation simulations for tropical cyclone events over Japan in a comprehensive manner.

Logan, Utah, Us

Washington, District Of Columbia, Us

I reviewed the proposals through a concept note criteria using an online platform. The proposals focused on (1) reusing wastewater, (2) capturing and storing water and (3) removing salinity of water supply in the developing countries.

Boulder, Colorado, Us

• Learnt Unidata software tools, such as: netCDF API, THREDDS, RAMADDA and IDV• Examined and recommended approaches for HydroDesktop to access Unidata online data repository. • Explored and recommended approaches to CUAHSI for using Unidata functionality to enhance the HIS cataloguing functionality that supports centralized search and data discovery.

Orono, Me, Us

Doctor Of Philosophy (Phd), Civil Engineering

Master'S Degree, Civil Engineering

Bachelor'S Degree, Civil Engineering