Hire Pushpum K. - AI Engineer

* Project 1: Develop Credit Scoring Engine for all customers. * Project 2: To assess the Risk profile of current customers of MNC Bank (Motion Banking) through AI. Project 3: Assign short term credit limit using customer-based segmentation by unsupervised clustering for Buy Now Pay Later (BNPL) service for MotionPay (MNC Product). * ● Started from scratch. Lead multiple meetings with the Bank team, Credit Card team and Risk Management team to discuss different use cases and data required. * ● Worked on multiple use cases namely Buy Now Pay Later (BNPL), Credit Score Engine, Credit Risk Assessment of current customers. * ● Processed and cleaned data of around 3,50,000 customers in excel format of size 30 GB and from three different domains i.e., Credit History (Bank), SLIK (Indonesian Bureau) and Demographics. * ● Performed comprehensive EDA after fetching data from AWS S3 for both multivariate and univariate analysis. * ● Performed Feature Engineering, Feature Selection and Feature Extraction to select the best method with top * predictive features. * ● Performed ML classification modeling using various models. XgBoost Classifier had the best accuracy. * ● Performed Unsupervised Clustering for BNPL use case using data from MNC E-Commerce platform (AladinMall). * ● Wrote backend and API using Flask on PyCharm and performed thorough debugging for errors. * ● Performed Refactoring, Debugging of the scripts and Testing through Postman. * ● Build and Deploy Docker file as compressed image on AWS EC2 through terminal. * ● Completed POC on Graph Database (TigerGraph, Neo4j), A/B testing, Docker, AWS EC2 * ● Worked with Off-Shore Data Science teams from China and Singapore.

* ● Performed thorough EDA including multivariate and univariate visualization on the data provided by the Data Engineers team. * ● Merged and cleaned a large amount of data from various outlets of a top bakery company in the USA. * ● Used Heat map and Feature selection through tree models to select useful features for dimensionality reduction. * ● Performed Feature Engineering to ensure noiseless data for the model building. * ● Worked in building various predictive models with client data as Input Features and Sales of outlets as output. * ● Developed Model building framework and tried various Models like Linear Regression, Rigde Regularization, Lasso Regularization, RandomForestRegressor, XgboostRegressor. RandomForestRegressor had the best accuracy. * ● Best model was selected based on Adjusted R-squared, MSE and RMSE. * ● Performed Unsupervised Clustering for second use case to segregate customers using behavioral data.

* Placed positions in Software Development, Engineering, Network Engineering, Project Management, Business Analysis, Data Analysis, and other technical/professional positions, PeopleSoft positions. * Developed and maintained candidate pipeline through direct sourcing, referral solicitation, contact networking. * Scheduled and coordinated client/candidate interviews, preparing the candidate with interview etiquette and company information. * Handled full life cycle recruiting, interviewing to closing candidates, client/candidate interviews, reference/background checks, managed candidates on site.