Divyam Khanna Email & Phone Number

Jersey City, Nj, Us

Jersey City, New Jersey, United States

Jersey City, New Jersey, United States

• Developed two GLM models to quantify risk and predict insurance pricing for Verisk’s flagship product ISO Risk Analyzer, enhancing pricing performance by 19% compared to industry standards.• Engineered a custom Python library for product-specific data science workflows, boosting team productivity and shortening project delivery times by 20 hours by streamlining routine tasks.• Optimized large-scale big data and modeling pipelines on AWS using parallel processing, slashing processing times by 80%.• Developed a Docker-based deployment and dashboarding platform in one month, enhancing delivery of 12 pricing models to the actuarial team and boosting team efficiency and collaboration.• Envisioned a bias reduction algorithm and fairness metric for insurance pricing, enhancing equity across demographic segments, aligning with ethical AI practices.• Won the Verisk Way to Go award for exceptional performance and spearheading a highly visible project.

Los Angeles, California, United States

Research work under Prof. Hernan Galperin for the project "Digital Divide in California".• Work with geo-spatial data for mapping broadband availability and community institutions over California.• Perform spatial and statistical analyses to observe patterns relating to community characteristics, K-12 school boundariesand other demographics and create a dashboard using Tableau.• Experiment with machine learning techniques to create an estimation model which assigns weights to certain geographies.

Jersey City, New Jersey, United States

• Successfully worked on improving two of the biggest insurance products for the corporation namely ISO Risk Analyzer and ProMetrix.• Researched and developed client-customized GLM models for predicting risk and pure premium for commercial businesses in AWS.• Effectively completed multiple data related ad-hoc tasks by multiple teams over the course of the internship.• Led a team of five to create a framework for record currency of commercial property attributes like occupancy and construction by using NLP and geographical matching with 3rd party acquired datasets.

Los Angeles, California, United States

Machine Learning research supervised and led by Dr. Marcin Abram for the project "Emulating Quantum Dynamics with Neural Networks via Knowledge Distillation".• Improved and tested various recurrent architectures for emulating quantum dynamic systems via curriculum learning and time series prediction.• Worked with a Linux supercomputer to train large models and write scalable machine learning code.• One of the main contributors of the GitHub repository containing scripts for training, testing and hyperparameter tuning leveraging parallel processing and unit tests for clarity and scalability.• Reviewed and experimented with various machine learning literature regarding interpretability, explainability and generalization.

India

• Developed libraries for the project, Smart Mirror, deployed on the RaspberryPi.• Built Face Recognition and Speech Recognition libraries with Python, JavaScript and OpenCV.

Gurgaon, Haryana, India

Developed REST APIs for the backend of a File Hosting Web App using the SpringBoot Framework.

Vellore, Tamil Nadu, India

Member of Creation Labs in VIT working on data extraction and analysis for custom wearable technology.

Master'S Degree, Applied Data Science, 3.8/4.0

Bachelor Of Technology (B.Tech.), Electronics And Communications Engineering, 9.08/10.0

Education record