Radhika P

Developed and delivered an automated user-friendly data cleansing tool using Streamlit, enhancing data quality and workflow efficiency. Collaborated with the client for the requirements and assumed leadership of the team for a week during an emergency situation. Awarded Spot Recognition Award for the contribution to the project. Executed and delivered a market basket analysis POC for a retail recommendation system project. Developed an OCR algorithm using Python and OpenCV to extract text from confidential documents as a part of the project. Led and organized the weekly technical discussion for the team as a part of Learning and Development program.

Developed machine learning models to classify liver lesion characterization features (hyper/ hypo enhancement, capsule, washout and diameter) and generate LIRADS scores. Aggregated medical data from various sources and implemented data processing pipelines to clean, preprocess, and transform raw data for model development. Explored various feature engineering, feature extraction (Pyradiomics), feature selection (Boruta, SFS), synthetic data augmentation (SMOTE, ADASYN) and ML approaches (Random Forest, XGBoost, etc.,) for developing algorithms. Supported the C++ integration activities for LI-RADS by developing feature extraction methods and the machine learning models using OpenCV framework for easier integration. Deployed the developed models in internal platform for further validation.

Courses included Data Science, Machine Learning, Linear Algebra, Deep Learning, Cloud Computing, and Advanced Data Structures and Algorithms.