Infrastructure Architect

Frontier's AI capabilities run at scale on a multi-tenant SaaS platform. The infrastructure that underpins our model serving, agent orchestration, and data pipelines needs to be cloud-native, hardened, and cost-aware from day one. This role owns that foundation.

The Role

You own the infrastructure the AI organization is built on. You design the cloud-native foundation — compute, networking, CI/CD, observability, and the model-serving layer — that lets a distributed team ship AI capabilities safely and at scale on a multi-tenant SaaS platform.

You codify infrastructure as code, harden it against the security realities of a SaaS environment, and keep latency, reliability, and cost under control as our AI workloads grow. You set the operational standard the rest of the AI organization follows.

Tech Stack

• Cloud: AWS (ECS Fargate, networking, IAM)
• IaC: Terraform
• Containers & CI/CD: Docker, automated build/test/deploy pipelines
• Observability: Datadog, Logfire, Sentry
• Model serving & routing: vLLM / TGI, LiteLLM
• Database & cache: PostgreSQL, MariaDB, Valkey

What We're Looking For

• 5+ years in infrastructure / DevOps / platform / SRE, with architecture ownership.
• Cloud depth — AWS strongly preferred (ECS/Fargate, networking, IAM).
• Infrastructure as code — Terraform (or equivalent) as a default way of working.
• Containers and CI/CD — Docker and automated build/test/deploy pipelines.
• Observability — Metrics, tracing, and logging (Datadog and similar) wired in from the start, not bolted on later.
• Security — Hardening cloud environments, secrets management, least-privilege access, and the operational discipline a multi-tenant SaaS requires.
• ML / LLM serving infrastructure — GPU workloads, model serving, inference routing, and cost control are a strong advantage.
• Communication — Clear architecture notes, runbooks, and proposals for a distributed team.