Summary
Michael McCulloch is an LLMOps/MLOps engineer with nine years of experience building resilient cloud ML systems that prioritize minimal surfaces and long-lived reliability. He has architected and deployed large-scale GPU/TPU workloads (H100/H200) and productionized open-source LLMs across AWS, GCP and Azure, often optimizing for cost and durability. His background spans hands-on software delivery from trading-platform features at Morgan Stanley to directing RAG and compiler-like systems at startups, bringing both deep systems knowledge and product sense. Michael enjoys re-deriving solutions from first principles—applying insights from semiconductor physics and biochemistry to engineer emergent behavior on CUDA hardware. Based in Calgary, he combines practical orchestration skills (Terraform, Bash, cloud infra) with a philosophical focus on building systems that outlast human error.
9 years of coding experience