Sumit Anand Email and Phone Number

Data Scientist, Insurance Risk Classification● Conducted EDA, feature engineering, and PCA, and tried various sampling techniques to reduce the class imbalance.● Executed ML Models (logistic regression, random forest, XGB, and decision tree) on AWS Sagemaker to predict the risk levels (likeliness of claims) and achieved an accuracy score of 89%.● Implemented and deployed the best fit model as a web application to make predictions using EC2 and Elastic Bean Stalk.Data Scientist, AI Assisted Q&A Retrieval Chatbot● Created Watson Assistant Chatbot and included the search skill to retrieve the most probable answers which have a high confidence score.● Set up Watson Discovery to extract the answers from business documents for the queries by service reps that reduced the turnaround time by 50%.● Built and trained the custom machine learning model in the IBM knowledge studio by identifying entities/relationships, and domain concepts.● Improved operational efficiency by performing model relevancy training.

Food Recommender System Application● Followed Agile Scrum methodology as a Product Owner and a core AI Team member.● Developed and tested Hybrid Recommendation System for a Food Industry Client using Collaborative Filtering and SVD algorithms to conditionally recommend other menu items based on users Diary to increase order sales.● Created a complex formula for rating the dishes for client data contingent on the user profile as per USDA guidelines.● Improved and Optimized the recommender code by using Pythonic way, thereby increasing performance and accuracy by 50%.● Developed Azure Function to expose Recommender System as a service which made it easy to deploy and secure.● Maintained Git branches, Created CI/CD pipelines, and managed deployments using Azure DevOps.Solution Architect – Airline Defect Estimation ● Designed System Architecture and proposed Tech-stack that kick-started the application development.● Designed Data Flow Diagram (DFD) which was used as a baseline for all application services.● Created DB schema, data tables & optimized queries to reduce response time by 30%.● Developed a predictive model based on a regression algorithm to estimate the airplane defects using historical data that reduce the manual effort and achieved model accuracy of 93%.App Developer – Smart Queue for Transportation Industry● Wrote Views and Models in Python Django Rest Framework to logically represent station queues, train schedules for Queue based Transportation system.● Developed backend services for the Reward Redemption System of the application that will reward the customers based on the queues selected.● Optimized and redesigned SQL queries to reduce the overall latency of the website and to reduce response time.● Performed bulk and load testing of the application using Jmeter to analyze its performance.● Dockerized and deployed the application on the Heroku platform.

App Developer, Insurance Domain Client● Involved in the software development lifecycle of the application like requirement gathering, requirement analysis, design, implementation, and testing.● Developed Business Logic to check for insurance claims using Python on Django Web Framework.● Implemented database access using the code-first approach of database mapping classes (i.e. Model) through Django Object Relational Mapping (ORM).● Created a Django dashboard with a custom look and feel for end users after a careful study of the Django admin site and dashboard.● Designed and implemented Django URL pattern for routing to call Django View that implements the business logic on the Models before generating the respective data presentation

Selected as one of the Technical leads for Google Developer Student Club at Drexel University.Responsible for giving training on new technologies like AI, ML, GCP to students.