Overview : Responsible for deployment, CI/CD, scaling, GPU cluster orchestration, and monitoring.
Responsibilities :
- Deploy GPU workloads using Kubernetes
- Configure GPU scheduling
- Deploy Triton Inference Server and NVIDIA TAO toolkit
- Implement CI/CD for ML models
- Set up monitoring: ○ GPU utilization
○ Model latency
○ Throughput
- Build rollback and A/B testing pipelines
- Automate model retraining pipelines