NLP-Based Compliance Model Data Labeling & AI Training Lead (Scan Logic Tech Services)
Developed and maintained an NLP-based compliance model for analyzing legal documents against law firm time/cost entries. Focused on extracting and classifying key provisions from legal and financial text to automate validation and flag potential violations. Led prompt engineering initiatives for generative AI to optimize model performance and compliance accuracy. • Automated document annotation and entity extraction workflows using internal and proprietary tools (Python, Azure). • Implemented reinforcement learning and reward-punishment mechanisms to iteratively enhance model labeling accuracy. • Built pipelines for processing, labeling, and validating compliance data from diverse business process sources. • Reduced manual validation efforts by integrating data labeling automation across Appeals, Collections, Cash Receipts, and Accounts Payable departments.