Ercan Ermis

I’m Ercan Ermis. Senior cloud platform engineer based in the Netherlands. I write here about AI, LLMs, agents, and the engineering work it takes to ship them.

How I got here

The first computer in my life was an Amstrad with two 5.25-inch floppy drives, bought by my father in 1986 for his business. The real switch flipped in 1998, fourth grade, when my teacher installed Linux on one of the Windows 95 machines in our school computer lab and said “this is Linux, it is free software.” Then Pac-Man appeared on that black screen and I was done. Why does this thing work, how does it work, what else can I make it do. More than thirty years later I am still asking those questions and still at the keyboard. Hours in front of a computer are still where I feel most comfortable. The AI work on this site is a continuation of the same curiosity, just with a new set of primitives.

The posts on this site come from production systems and side projects, not slide decks. Cutting Bedrock Knowledge Base costs by 90 percent by moving off OpenSearch Serverless. The agentic primitives that landed at re:Invent. What Kiro changes for the dev loop. The pattern under S3 Vectors. Each post is what it actually looked like to do the work, with the numbers and the failure modes attached.

Where this fits

AI moves fast. The hype layer moves faster. I write for engineers who already know what an embedding is, who have shipped an LLM call to production, and who want the trade-off, the cost number, or the failure mode that did not make it into the launch keynote.

The bias is toward:

  • Real numbers from real workloads, not synthetic benchmarks.
  • Cost-aware patterns. Token spend, vector storage, inference latency all matter.
  • Honest failure-mode analysis. What breaks at scale. What does not survive a Monday.
  • Engineering hygiene applied to AI. Same review, same observability, same blast-radius thinking.

Credentials and community

Projects

  • news.ercan.ai, Curated AI news, link drops, and opinionated bulletins. Short-form feed; this site (ercan.ai) is for long-form field notes.
  • UptimeCoach, Personal SaaS lab for global uptime monitoring, provisioned across 20 AWS regions as a working multi-region reference architecture.
  • awsmonthly.cloud, Monthly AWS news magazine. Digest-style roundup of what shipped, what changed, and what matters. (Coming soon, haven’t had time yet.)

Other writing

Consulting and advisory

I take on a small number of AI and applied-ML consulting engagements. I like this work. It keeps me inside real production constraints, which is where the useful patterns come from.

Ways I work with teams:

  • AI/ML platform advisory. Your team is moving beyond the prototype phase and needs a cost-aware, secure foundation for LLM workloads. Bedrock, SageMaker, vector storage, inference routing, guardrails. I help design the architecture and the operating model around it.
  • Project-based build. You have a specific AI feature or pipeline you need shipped. I build it with your team, transfer ownership, and document the failure modes so you are not rediscovering them next quarter.
  • AWS cost optimization for AI workloads. Bedrock Knowledge Base, OpenSearch Serverless, provisioned throughput. Most teams overpay by 60-90% because the pricing model is non-obvious. I find the waste, restructure the stack, and hand you the cost dashboard.

The broader surface I consult on, in case the shape of your problem is wider than AI: AWS, cloud architecture, CI/CD pipelines, Linux, GitHub and GitLab tooling, Terraform, Kubernetes and EKS, observability, cost optimization, and migration. The AI workloads sit on top of that stack; in practice the two are one engagement, not two.

These are small, focused engagements. If what you are building overlaps with the topics on this site, reach me on LinkedIn. Tell me what you are trying to ship and what the bottleneck looks like.

Read this next