AI/ML Platform Engineer — Hybrid Cloud & GPU

Deepgram; United States; Hybrid; USD 180,000 - 260,000; Full time

Job summary

Deepgram is seeking an experienced Platform Engineer to build and operate the hybrid infrastructure foundation for our AI/ML research and product development. You will architect, build, and run the platform spanning AWS and on‑prem data centers, empowering our teams to train and deploy complex models at scale.

This role emphasizes self‑service environments, Kubernetes, Terraform, and GPU scheduling with Slurm, plus on‑prem bare metal and a focus on observability and automation.

Qualifications

• 5+ years of Platform Engineering, DevOps, or SRE experience.
• Hands-on production infrastructure with Terraform.
• Expert knowledge of Kubernetes architecture and operations at scale.
• Strong scripting and automation (Python, Go, Bash).
• Experience with CI/CD systems and developer tooling.

Responsibilities

• Architect and maintain our core computing platform using Kubernetes on AWS and on‑premise.
• Develop and manage infrastructure using Terraform to ensure reproducibility and automation.
• Design and optimize AI/ML job scheduling and orchestration with Slurm and Kubernetes.
• Provision, manage, and maintain on‑premise bare metal server infrastructure for GPU computing.
• Implement networking and storage to support hybrid workloads.
• Develop observability, monitoring, logging, and automation for operations.
• Collaborate with AI researchers to build tools and workflows that accelerate development.

Skills

Kubernetes

Terraform

Slurm

AWS

Python

Go

Bash

CI/CD tooling

Tools

Terraform

Kubernetes

Slurm

AWS

GitLab CI

Jenkins

ArgoCD

Job description

Deepgram is seeking an experienced Platform Engineer to build and operate the hybrid infrastructure foundation for our AI/ML research and product development. You will architect, build, and run the platform spanning AWS and on‑prem data centers, empowering our teams to train and deploy complex models at scale.

This role emphasizes self‑service environments, Kubernetes, Terraform, and GPU scheduling with Slurm, plus on‑prem bare metal and a focus on observability and automation.