Issue 4
Weekly insights from an AI engineer building real systems.
A small packaging mistake exposed Claude's source — a reminder that most incidents aren't sophisticated, just preventable.
In this issue: what went wrong, why it matters, how it reflects our growing reliance on AI, and a checklist you should probably be using.
Skills in Google ADK aren't just a pattern — they're how you stop agents from hallucinating and start building something you can actually ship.
Read the article →Mini insights — quick, unfiltered thoughts I don't publish anywhere else.
Subscribers only.
Hi, I'm Ayushi.
I'm an AI engineer working on LLM-powered and agentic systems, and still learning something new about this space every day.
Over the past few years, I've spent a lot of time building, experimenting, and trying to understand what actually works beyond demos and POCs.
This newsletter is where I share what I'm learning — the useful bits, the mistakes, and the things that feel worth paying attention to.
If you'd like to follow along, you can subscribe and get this in your inbox.