Hi, I’m Shawn Azar.
I’m a technologist based in Denver, Colorado, and I’m obsessed with tinkering.
When I’m not spending time with my beautiful partner and family or getting outdoors, you’ll find me in my homelab experimenting with GPU clusters for LLM inference, optimizing Kubernetes deployments, or testing the latest CNCF projects to see which ones actually deliver on their promises.
I’ve spent the last decade working with fast-growth startups, helping organizations like Akerna, Bonusly, and Moment navigate the technical challenges of going public or raising capital. That work taught me how to build systems that scale under pressure, how to make architectural decisions when there’s no time for perfect, and how to translate complex technical problems into business outcomes.
Currently, I’m a Senior Engineering Manager of Foundation at Strava, where I lead infrastructure initiatives. Before that, I managed CloudOps teams and built scalable cloud solutions that consistently delivered 40-60% cost reductions while improving reliability.
The last five years, I’ve been deep in cloud architecture, machine learning, and AI/agentic systems. I’ve architected gaming server infrastructure with Kubernetes and Agones.dev, built asset management services that slashed deployment times, and designed RAG systems that perform reliably at enterprise scale. My homelab has become a testbed for practical AI infrastructure, from local LLM inference on custom GPU setups to evaluating emerging tools in the agentic workflow space.
What I write about here:
This site is where I share what I’m learning as I build and tinker. You’ll find posts about:
AWS cost optimization at scale: Lambda functions, RDS databases, S3 storage, CloudWatch, data transfer, and Reserved Instances. Real tactics that drive 40%+ savings without sacrificing performance.
AI/ML infrastructure: Building production-ready RAG systems, LLM observability from token-level metrics to reasoning traces, deciding between fine-tuning and retrieval augmentation, and optimizing AWS costs for machine learning workloads.
Cloud architecture patterns: Designing resilient, cost-effective systems that scale. High availability strategies, microservices architectures, and infrastructure migrations from on-prem to cloud-native.
Kubernetes and CNCF ecosystem: From Agones.dev for gaming servers to emerging projects like kagent. What works in production versus what’s just hype.
Agentic workflows and MCP servers: Practical experience building AI tools that handle complex business processes. Which MCP servers are production-ready (Cloudflare, Hostinger, WordPress, GitHub, Kubernetes), compatibility nightmares, and token cost realities.
Infrastructure engineering leadership: Managing CloudOps teams, building ITSM systems, implementing monitoring with Prometheus and Grafana, and creating CI/CD pipelines that actually improve developer velocity.
Homelab experiments: GPU setups for local LLM inference, infrastructure testing, and the kind of tinkering that informs production decisions.
I write because I wish more people shared the real experience of building with these technologies: the 3am incidents, the cost optimization discoveries, the architectural decisions that looked brilliant on paper but failed in production. The actual work, not the conference talk version.
If you want to work together:
I consult on AWS architecture, AI infrastructure implementation, Kubernetes management, and cloud operations that need to work reliably at scale. If you’re building something that needs to handle growth, dealing with runaway cloud costs, or want to integrate AI tools without the hype cycle, let’s talk.
Through High Country Codes, I’ve delivered CloudOps as a Service, security solutions, and cost optimization strategies for businesses of all sizes since 2010.
Connect:
Want to discuss a project, compare notes on infrastructure challenges, or just talk about the latest CNCF sandbox project? Reach out through my contact page.
You can also find me on LinkedIn, GitHub, and occasionally Strava (when I’m not in the homelab).