Beyond Vibe Coding: Building Reliable AI AppSec Tools with Emily Choi Greene

Security teams are overwhelmed by AI‑generated code, but how do you actually use LLMs for real application security? In this BSides Vancouver Island talk, Emily Choi‑Greene (CEO & Co‑Founder, Clearly AI) explains how to turn large language models (LLMs) into reliable, production‑ready AppSec tools. Drawing on her experience securing Amazon Alexa and building AI security platforms, she breaks down practical strategies for using AI in cybersecurity, secure coding, and software security workflows. This session is ideal for AppSec engineers, security architects, blue teams, DevSecOps practitioners, and engineering leaders exploring AI security tools. Key topics include: - Using LLMs for application security and secure code review - Why RAG (retrieval‑augmented generation) often fails in security use cases - Reducing hallucinations with structured outputs, source citation, and LLM‑as‑judge - Designing AI systems for high reliability (“multiple nines”) - Combining agents, tools, a