← All work

Infrastructure / Technical leadership

ML Infrastructure at Snap

Continuous ML, labeling, training, deployment, and GPU infrastructure for camera intelligence and gen AI.

MLOps architecture diagram from KubeCon presentation

My contribution

  • Grew the ML infrastructure team from 0 to 10 and defined team systems for hiring, career growth, and execution.
  • Presented continuous ML with Kubeflow and Spinnaker at KubeCon and training-data pipelines at AWS re:Invent.
  • Worked on GPU infrastructure and fine-tuning job scheduling for Snapchat Dreams.

MLOps / Kubernetes / Kubeflow / Training data