Building Secure LLM Workflows in Production
Lessons learned architecting GenAI systems that are actually safe to ship — from prompt injection mitigation to output validation and observability.
Read post Writing
Blog
Notes, opinions, and deep-dives from over 10 years on the stack. I write about what I'm actually building — not what sounds good in theory.
Apache Beam and the Art of Scalable Pipelines
A practical look at what makes Apache Beam's unified model so powerful — and the gotchas you'll hit when you take it to production at Google scale.
Read post