Supporting the Finance Automation team in enhancing operational efficiency and elevating customer satisfaction through the development and implementation of machine learning (ML) solutions:* Built Entity Extraction Service leveraging a unique Alpha Num masking technique, alongside Bi-LSTM and CRF layers. This innovation facilitated the automated matching of 85% of emails to corresponding customer accounts.* Expanded the capabilities of the Entity Extraction Service to encompass the extraction of crucial information from invoices, achieving an impressive 96% First Pass Yield.* Developed a Proof of Concept (POC) for Zero Touch Response (ZTR), enabling automatic responses to inbound customer emails through the utilization of fine-tuned Large Language Models (LLMs).* Built Intent Classification Service utilizing a Masked Language Model (MLM) approach, fine-tuning it with custom datasets on RoBERTa architecture. This service achieved a remarkable 98% F1 Score in multi-label classification of emails.* Submitted 2 research papers to Amazon Machine Learning Conference (AMLC) 2023 to share findings with Amazon Science community.