Vamshi D.

Senior Nlp Engineer @Bny Investments Mellon

New York, NY, US

Architected and delivered end-to-end MLOps governance frameworks utilizing MLflow and Kubeflow Pipelines, fully automating model retraining and maintaining 99.7% system uptime for client deliverables.• Designed and provisioned Kubernetes-native model serving infrastructure on AWS EKS, configuring dedicated GPU node pools and HPA autoscaling to support 10K+ RPS across production ML services.• Engineered and deployed BERT Transformer NLP pipelines for complex customer sentiment analysis, serving them as REST microservices on AWS SageMaker to process 1M+ interactions daily.• Optimized inference latency for mission-critical delivery date prediction models (LSTM/GRU ensembles) through model quantization, successfully reducing response times from 500ms to 50ms.• Implemented robust automated drift detection mechanisms (PSI) and integrated them with monitoring dashboards to trigger continuous model retraining workflows without manual intervention.• Orchestrated complex CI/CD deployment pipelines using GitOps methodologies (ArgoCD) and Istio-managed canary rollouts to safely transition newly trained deep learning models into production.• Developed comprehensive logging and telemetry architectures using Prometheus and Grafana, tracking real-time SLA metrics on model accuracy, latency, and throughput for stakeholder visibility.• Collaborated directly with cross-functional data engineering teams to translate business requirements into scalable, containerized AI architectures.Environment: Python, FastAPI TensorFlow, PyTorch, Scikit-learn, Hugging Face (BERT/GPT) MLflow, Kubeflow Kafka, PySpark Docker, Kubernetes (EKS), Helm, Istio, ArgoCD AWS (SageMaker, Lambda, S3, EC2) Tableau, Power BI.

Jawaharlal Nehru Technological University Hyderabad (JNTUH)

Bachelor of Technology - BTech, Computer Science