Arjun Rao

Built enterprise GenAI/RAG pipelines using Gemini APIs, LangChain, Vertex AI, and Pinecone/semantic search to improve document retrieval accuracy by 38%. Deployed GenAI workflows for contextual summarization and intelligent search, reducing manual research effort by 45%. Improved recommendation quality by 29% using transformer-based NLP models. Processed 20M+ records daily with distributed PySpark/Kafka for anomaly detection and forecasting with near real-time inference. Reduced inference latency by 32% via Kubernetes autoscaling, GPU optimization, and quantization. Automated deployment/retraining/monitoring using MLflow, Kubeflow, Jenkins, and CI/CD, reducing manual effort by 40%. Developed low-latency FastAPI/Flask microservices on GKE and collaborated with cross-functional teams to improve delivery and reliability.

Improved forecasting accuracy by 27% for customer analytics, operational intelligence, churn prediction, and reporting automation. Reduced manual document processing effort by 41% through NLP solutions for sentiment analysis, entity extraction, and text classification. Built Spark and Airflow ETL pipelines for structured/unstructured data and scalable feature engineering. Accelerated deployment and batch inference by 33% via cloud-native ML integrations across AWS and Google Cloud. Increased model precision by 24% using feature engineering, hyperparameter tuning, preprocessing optimization, and continuous evaluation. Developed Tableau/Power BI dashboards for KPI monitoring and model performance visibility, and supported model governance with monitoring, drift analysis, and retraining workflows.