Before launching our AI assistant, we worked with a red teaming vendor (let s call them L ) to check how safe our product really was.

We were expecting a few corner cases or prompt injection attempts.

What we got was a pretty eye-opening report: infinite output loops, system prompt leaks, injection attacks that bypass moderation, and even scenarios where malicious content could be inserted by users via email inputs.

Hey everyone! I m Soomin, a Product Manager based in Seoul.

Over the past few years, I ve worked on various AI-driven and mobile services, and I m currently building a personal AI assistant that helps people get things done more effortlessly.

Lately, I ve been exploring how emerging LLM technologies and services can be meaningfully integrated into products especially to narrow the gap between intention and execution.

I love connecting with folks who are passionate about the same space.