Senior Infrastructure Engineer (ML Inference)

About the role

Artificial Agency, a player-facing agentic AI company for games, is looking for a Senior Infrastructure Engineer to build, operate and scale the inference services that make its AI agents responsive, reliable and cost-effective. Reporting to the Head of Platform Engineering, you will join the team running a multi-service, multi-cloud Kubernetes platform, focusing on GPU-backed inference across clouds and managed providers. The role is based in Edmonton, Alberta, hybrid, with remote possible for exceptional candidates.

Key responsibilities

• Build and operate GPU inference infrastructure, including scheduling, capacity planning, autoscaling and warm-up strategies.
• Evaluate AWS, GCP and managed providers against workload requirements, availability, reliability and cost.
• Share ownership of platform reliability, scaling, security and spend, including Kubernetes, Terraform, Helm, ArgoCD, PostgreSQL and S3.
• Partner with security and compliance colleagues on access controls and SOC 2 / ISO 27001 requirements.
• Benchmark agent workloads with backend and ML engineers and optimize latency, throughput, caching and memory use, guiding hardware selection and capacity purchases.
• Build repeatable deployment, validation and rollback workflows for models, adapters and serving configurations.
• Implement health-aware routing and provider failover, including timeouts, retries and overload handling.
• Monitor request latency, errors, queueing, GPU health and inference economics, maintain alerts and runbooks, take part in on-call and resolve incidents.
• Write and review production code and infrastructure configuration, automate operations, test deployments and recovery changes, and keep technical documentation up to date.

Required profile

• 5+ years building and operating production cloud infrastructure, with ownership of uptime, incidents and cost.
• Strong production Kubernetes experience and infrastructure-as-code with Terraform and Helm, and GitOps with ArgoCD or similar.
• Hands-on experience deploying, configuring and troubleshooting GPU-backed inference workloads; production LLM-serving experience strongly preferred.
• Experience load testing and diagnosing performance bottlenecks, including latency, time-to-first-token, throughput, batching and KV-cache trade-offs.
• Strong AWS and/or GCP networking and security fundamentals (IAM, secrets management) and production observability with Prometheus/Grafana, OpenTelemetry or similar.
• Proficiency in Python, sound testing and code-review practices, and collaboration across platform, backend and ML teams in a fast-moving startup.

Required skills

• Kubernetes, Terraform, Helm, ArgoCD, Python
• AWS and/or GCP networking and security
• Prometheus/Grafana, OpenTelemetry
• GPU scheduling, vLLM, TensorRT-LLM, Triton, quantized models, LoRA adapters, managed inference providers or multi-provider routing (valuable)