1) Architected and deployed scalable databases and ETL pipelines on PostgreSQL to support real-time machine learning model deployment, enhancing system performance by 27%.2) Enhanced model accuracy by 22% through prompt engineering and fine-tuning OpenAI’s Large Language Models (LLMs).3) Coordinated Power BI dashboards integration with backend APIs, improving stakeholder engagement and data-driven decision-making across teams.4) Innovated and deployed GenAI-based models, revitalizing search and retrieval systems through RAG optimization.5) Led the implementation of Redis server, enabling multi-user communication for enterprise software through isolated rooms.6) Developed secure RESTful APIs for seamless integration of machine learning models and automated deployments with CI/CD pipelines.

This project was under Microsoft AI for Earth where I secured a $15,000 Research Compute Grant from "Microsoft AI-For Earth" for my project, "AI-Based Statistical Analysis of Pollution Due to Land and Ocean Plastic Waste in India." Aligned with Sustainable Development Goal 13, our objective was to estimate plastic pollution's impact in India, encompassing land and water bodies leading into the ocean.Faced with limited data, I orchestrated data compilation from diverse sources to construct a comprehensive database. Leveraging machine learning, we predicted plastic pollution impact from both terrestrial and aquatic sources into the ocean. We overcame data limitations through advanced multivariate regression techniques, enhancing predictive model accuracy.For dynamic data representation, Power BI was pivotal, facilitating insightful visualizations. This multidisciplinary approach, blending advanced statistics and machine learning, advances our fight against plastic pollution. #AIforEarth #DataScience #EnvironmentalSustainability 🌍♻️