● Integration of Advanced Methodologies in LLMs:Adaptation Using FLAN and ORCA: Employed advanced methods from FLAN and ORCA to create highly effective fine-tuning datasets. ● Domain Adaptation with QLoRa Methods: Utilized LoRa and QLoRa methodologies to adapt LLMs to specific sub-domains.● Development of a Self-Correcting Agent: Innovated within the ReAct framework to create a self-correcting agent that induced logical reasoning,● Enhanced Q&A Retrieval: Worked on Q&A retrieval using RAG and advanced filtering systems, achieving a 40% improvement in information relevance and retrieval speed. ● DETR Model Fine-Tuning for Document Analysis:Specialized in adapting the DETR model to the nuanced requirements of document analysis, improving its ability to identify and categorize textual and non-textual elements within complex documents. ● Architecture Augmentation for Enhanced Object Detection:Led a project to augment the LayoutLMV3 architecture, specifically targeting the enhancement of its object detection capabilities. ● Authored and Published Articles on Explainable AI: Spearheaded the creation of engaging and informative content aimed at demystifying AI principles and practices for the general public. ● My articles, published on behalf of Walnut.ai, have:1. Attracted significant attention from industry professionals and media, increasing our company's visibility and influence in the AI space.2. Contributed to a 20% increase in engagement with our company’s digital content platforms, highlighting the growing interest and appreciation for accessible AI knowledge.

Statistical Modeling for Insurance:Developed statistical models resulting in a 95% true positive rate for early claim detection, reducing false claim costs by over 30%.Enhanced predictive accuracy, leading to a 25% increase in operational efficiency for the claims processing department.