Architect and deploy scalable AI infrastructure for LLMs and multi-modal models—covering data pipelines, model serving, and inference optimization on Kubernetes/AKS/EKS.
Design GPU/NPU-aware cluster topologies with auto-scaling, model checkpointing, and low-latency inference SLAs for production workloads.
Integrate MLOps toolchains (MLflow, Weights & Biases, KServe, Triton) with CI/CD pipelines to automate model deployment, rollback, and drift detection.
Establish infrastructure-as-code (IaC) standards using Terraform or Pulumi for reproducible, secure, and auditable cloud environments.
Collaborate with ML Engineers and Data Scientists to optimize model quantization, batching strategies, and memory footprint for production efficiency.