Robert Belson

About Robert Belson

As a distributed systems engineer and solutions architect, I design and build production-scale AI and edge computing infrastructure at Amazon Web Services (AWS). My work centers on deep infrastructure challenges: from GPU orchestration across distributed Kubernetes clusters (EKS on Outposts, Hybrid Nodes, Local Zones) to architecting hybrid Retrieval-Augmented Generation systems that span cloud and edge. I specialize in the foundational layer where networking, compute scheduling, and distributed AI inference converge, enabling startups and enterprises alike to deploy latency-sensitive workloads at planetary scale.With a Master\'s in Computer Science focused on Networks and Distributed Systems from Northwestern University, I\'ve contributed to academic research on web performance optimization and hold three patents in cellular network energy modeling, location-based access control, and geographic traffic restriction systems. Prior to AWS, I led Verizon\'s Developer Relations team focused on mobile edge computing nd conducted distributed systems research at Northwestern\'s AquaLab, focusing on network optimization in challenging environments.I\'m passionate about solving hard infrastructure problems — whether that\'s optimizing network packet flows, designing resilient distributed architectures, or enabling GPU workload orchestration across hybrid cloud environments — and sharing these solutions through open-source contributions or AWS On Air, AWS\' flagship weekly show where we reach millions of developers each year.

Experience

1. Senior Solutions Architect Generative Ai Innovation & Ecosystem Development

   Amazon Web Services (AWS)

   Feb 2025 — Present

   Architecting and implementing scalable generative AI solutions, with focus on telecommunications and enterprise verticals. Leading solution engineering, open-source contributions, and technical evangelism across strategic AI infrastructure patterns, including:• Hybrid Retrieval-Augmented Generation (RAG): Designed and open-sourced a production-grade Hybrid Retrieval Augmented Generation pattern integrating Amazon Bedrock with EC2 Edge infrastructure. Delivered as reference architecture, hands-on workshop and customer implementation artifacts.• Kubernetes GPU Orchestration at the Edge: Led engineering effort to enable native NVIDIA Run:ai integration across distributed Amazon EKS deployments— spanning AWS Local Zones, Dedicated Local Zones, EKS on Outposts, and EKS Hybrid Nodes. Enabled GPU workload management and scheduling for edge AI/ML workloads.Technical Focus Areas:• Distributed AI inference architectures• Edge computing and hybrid cloud patterns• GPU orchestration and resource management• LLM deployment and optimization

Education

• Northwestern University

  Bachelor of Arts - BA, Computer Science

  2015 — 2019

• Levine Music

  Classical Guitar Performance

  2005 — 2015

• Northwestern University

  Master of Science - MS, Computer Science

  2018 — 2019

• Northwestern University

  Bachelor of Music - BM, Classical Guitar Performance

  2015 — 2019

Skills

• Mobile Advertising
• Entrepreneurship
• Classical Guitar
• Business Strategy
• Software Development