Gayatri Devi Dasari

Collaborated with senior ML engineers to design and deploy BharathHealthAI, an AI-powered healthcare assistant integrating LLaMA, Retrieval-Augmented Generation, and multi-agent orchestration, improving patient query resolution speed by 45%. Supported semantic search integration using vector databases for medical literature retrieval, enhancing diagnostic accuracy. Built and fine-tuned 12 disease prediction models using Scikit-learn, TensorFlow, and PyTorch CNNs for early screening of diseases including diabetes and various cancers. Developed FastAPI-based AI agent endpoints for multi-step healthcare workflows, contributing to broader platform adoption among healthcare providers.

Contributed to HealthAI, an AI-powered healthcare assistant integrating LLaMA, RAG, and multi-agent orchestration, improving patient query resolution speed by 45% in pilot deployments. Integrated semantic search with vector databases to retrieve medical literature and case studies, enhancing diagnostic support and clinician trust. Built and fine-tuned 12 disease-prediction models (diabetes, heart disease, cervical and throat cancers) using Scikit-learn, TensorFlow, and PyTorch CNNs for early screening from structured health records. Developed FastAPI-based AI agent endpoints to manage multi-step workflows and context-aware reasoning, deployed on GKE for scalable partner adoption.

Collaborated with analytics teams to analyze over 10 million customer transactions to identify churn trends and support retention campaigns, reducing churn by 8%. Built clustering and supervised models to rank top at-risk customers for targeted marketing. Contributed to an anomaly detection workflow that reduced false positives by 20% and improved fraud detection by 15%. Played a key role in implementing TensorFlow and PyTorch CNN-based object detection models for automated product defect detection, reducing inspection time by 30%. Assisted in setting up CI/CD automated ML pipelines and model version control to ensure consistent production model updates. Developed ETL pipelines to clean and transform data for analytics and maintained KPI dashboards in Power BI, Tableau, and Plotly for strategic business reviews.