Your LLM Is Lying to You and You Have No Idea
I spent three months building LLM-powered features before I realised I had no idea what my models were actually saying in production. Here is what I found — and what I wish I had known from the start.
Written by
I keep systems alive at scale. Now I write about how.
SRE with 8+ years of experience across AWS, Azure, and GCP. I've built multi-cloud Kubernetes platforms from the ground up, designed CI/CD pipelines, and owned high-pressure incidents at scale — including keeping systems alive during Walmart's Black Friday. I care about making on-call less painful — through better alerting, smarter observability, and automation that actually works. Currently on a 30-day public learning challenge — writing one blog every day about what I'm relearning and rebuilding. Topics: Kubernetes, Prometheus, Terraform, Python, Go, system design, and everything in between. Follow along if you're on a similar journey.
Responses
Markdown supported · ⌘↵ to post
No responses yet.