Aasritha Bhimisetty

Led the development of an AI-powered real-time underwriting and adaptive pricing system, delivering personalized risk insights and improving efficiency. Built models (XGBoost, LightGBM, LSTM) for claim frequency, churn, and time-series forecasting; applied CNNs (ResNet-50) for image-based claims to automate damage detection. Created end-to-end MLOps pipelines with Kubeflow Pipelines, Vertex Pipelines, and Airflow for scalable data processing. Implemented microservices (FastAPI/Flask) for low-latency inference and integrated with enterprise APIs. Deployed models on Google Vertex AI and AutoML, and maintained vector databases (FAISS, Pinecone, Weaviate) for RAG-driven reasoning. Fine-tuned open-source LLMs (LLaMA, Falcon, Mistral) and used Gemini/Claude APIs for domain adaptation in insurance tasks. Orchestrated distributed systems with Kubernetes/Docker, enabling real-time fraud detection and underwriting predictions. Built RAG architectures and semantic knowledge graphs with Neo4j and

Led AI/ML initiatives across NLP, sentiment analysis, and demand forecasting. Implemented dynamic percentile scoring integrated with Tableau dashboards to monitor recurring customer service incidents in real time. Built and deployed ML models (Random Forest, GBM, SVM) for demand forecasting, inventory optimization, and logistics. Integrated multiple data sources to create a 360-degree customer view and supported real-time decision-making with Apache Kafka. Built scalable Dockerized models and used Hadoop/PySpark in AWS for large-scale analyses. Pioneered Generative AI initiatives with GANs for synthetic data to bolster model robustness. Explored GNNs for enhanced fraud detection and implemented CNNs for image-based product recognition. Worked on real-time streaming, IoT analytics, and cloud-based model management with BigQuery and Azure ML. Implemented clustering (K-means, PCA) for customer segmentation, and performed hyperparameter tuning via GridSearch/RandomSearch. Led store analyti