Finding Deduplicates in the record system: Developed a framework for identification of redundant customer records in the system, multiple contact information associated for the same customer. The manual curation process which is ineffective and time consuming is replaced by AI-based techniques so the capability to effectively communicate with the customers to upscale the cross selling / up-selling opportunities is achievable

Anomaly Detection of Mixed Attributes: Based on a research paper, built a POC model for handling Numerical and Categorical attributes for Anomaly Detection in the dataset. This model was used on analysis of fraudulent behaviour in financial sector where detecting fraud in early stage reduced financial and reputation losses