About

Romee Panchal is an accomplished Data Scientist with over 7 years of experience solving complex problems across various industry verticals such as banking, finance, insurance, retail, and manufacturing. He is well-versed in machine learning, statistical analysis, deep learning, and natural language processing. Romee has a proven track record of developing high accuracy ML models and building efficient applications. He is experienced in writing clean, optimized, and readable code. Romee's expertise in churn, propensity, recommendation, forecasting, credit risk, anomaly detection, classification, and regression problems has helped organizations make informed decisions. Currently, Romee is working as a Data Scientist at PayU where he is solving finance problems. He has developed a graph pipeline to score users at scale, used PySpark to process raw data, and deployed docker containers for feature creation and model scoring using AWS Fargate instances. He has also developed a model monitoring framework that enables a new model to be added to the framework by updating the configuration file and feature creation steps. Romee has worked on a risk model for inactive users as well. In his previous role as a Data Scientist at Tiger Analytics, Romee developed personal loan propensity models for first-time buyers and repeat buyers. He also worked on early warning models for credit risk to identify defaulters in advance. Romee has built an end-to-end machine learning pipeline and deployed the solution in production. He built a model monitoring framework to regularly evaluate how the production model is performing. As a Senior Analyst at Tiger Analytics, Romee built anomaly detection and forecasting models for preventive maintenance for the manufacturing industry. He developed a generic model framework for anomaly detection and forecasting so that models can be quickly built for the same family of equipment. Romee designed and built an application to crawl data from various sources in a parallelized way and store the same in structured format for analytical usage. Additionally, he developed libraries for common functions to perform address matching, name matching, address formatting, calculate geo-distance between addresses, etc. This library helped others save a lot of time. Romee holds a Bachelor of Technology degree from Gurukula Kangri University. He has a strong tech stack that includes Python, AWS, Docker, Pytorch, SQL, Seaborn, OpenCV, Scikit, and Pandas, among others.