Greg Coquillo

About

I build AI systems at scale. This includes models, infrastructure, products, and operational systems required to take AI from prototype to production. Over the past 6+ years across Amazon and AWS, I led product initiatives at the intersection of AI, data, and large-scale distributed systems. I drove the adoption of AI and ML across Amazon Retail through media automation and decision intelligence, while at AWS I focused on the software layer that enables global AI network infrastructure. My work has centered on one core problem: how to operationalize AI reliably across complex, high-scale environments. Today at Microsoft’s Azure AI & HPC, I am focused on one of the most critical challenges in the industry: building the foundation for the next generation of AI. I am leading world-class technologists under Fairwater to deliver AI data center systems designed to support hundreds of thousands of NVIDIA GPUs, millions of High Performance Computer (HPC) vcores, powering frontier models for customers such as OpenAI, Anthropic, and Microsoft AI. This is where product, infrastructure, and physics converge. Power, networking, latency, and reliability are no longer backend concerns. They are the product. I do think that, in addition to better prompts, for AI to scale it would also require better systems. That includes: • High-performance compute and GPU orchestration • Distributed data and retrieval architectures • Production-grade AI workflows with observability and cost control • Tight integration between software, firmware, hardware, and energy systems Beyond my core role, I am a two-time LinkedIn Top Voice in AI and Technology, where I share practical insights on building real-world AI systems. I focus on helping technical leaders and builders understand how to move from experimentation to production with clarity and rigor. I invest in and advise early-stage AI companies through Leading Edge Ventures, with a focus on teams solving real problems with durable technical advantages. If you are building in AI and thinking about scale, infrastructure, or production systems, we are likely working on similar problems. (Opinions are my own)