- Gathered Data from Business, Analyzed and provided valuable insights for improving decision making :I have been actively engaged in collaborating with business stakeholders to collect and compile vast amounts of data, which often consisted of millions of records. To ensure accuracy and usability, I undertook a comprehensive analysis of this data, leveraging tools like Microsoft Excel to clean and refine the information.By utilizing advanced features such as charts, pivot tables, DAX (Data Analysis Expressions), Power Query, and Business Intelligence (BI) tools, I was able to extract meaningful insights from the data.These insights have proven invaluable in guiding decision-making processes and providing valuable information to support business strategies and initiatives.- Implemented POC for Fortune 4 client :I had the opportunity to work on the Google's Large Language Model (LLM) T5 model, utilizing fine-tuning techniques to customize its performance for predicting ICD codes. My work involved carefully tuning the model's parameters and optimizing its architecture to achieve the desired output through meticulous testing and iterative improvements, I successfully attained an impressive accuracy rate of 92%, showcasing my ability to deliver precise and reliable results in complex healthcare data analysis. With a passion for cutting-edge technology and a strong commitment to achieving excellence, I am eager to contribute my expertise to challenging projects in the field of natural language processing and healthcare AI.

- Developed POC for largest HealthCare provider of United States.As a professional working on Azure Cognitive Search, I focused on enhancing the ranking mechanism for patient searches on a website as per client's requirement. This involved implementing various strategies to improve search results and provide more relevant information to users. By utilizing Azure's powerful relevance and scoring algorithms such as BM25 (Best Matching 25) and Classic Similarity algorithms, I optimized the search process by considering different criteria, including relevance and context. Additionally, I introduced fuzziness & boosting parameters to handle variations in search queries and ensure more accurate results and achieved accuracy upto 96%. The ultimate goal was to enhance the user experience and make it easier for healthcare provider's to find the information they needed efficiently.