Vamshiteja Kaspe

Built Python and XGBoost fraud risk models with feature engineering, evaluation, and iterative tuning. Developed graph-based fraud detection workflows and LLM-based contextual explanation using LangChain with RAG. Implemented retrieval-focused AI flows, data pipelines, and REST services, and deployed ML components with Docker/Kubernetes. Added model monitoring and MLOps automation using SHAP interpretability, MLflow tracking, and retraining workflows.

Developed anomaly detection and cloud-cost forecasting models using Python with Scikit-learn and TensorFlow. Built multi-region telemetry pipelines with SQL and Airflow, and implemented production RAG and agentic workflows using LangChain and LangGraph. Delivered FastAPI services and supported Azure ML ingestion, training, validation, and deployment, improving CI/CD reliability. Created Tableau dashboards and contributed to troubleshooting and workflow documentation.

Developed an AI computer-vision approach for detecting face masks using TensorFlow. Focused on building the vision model and validating it for face-mask recognition use cases.