AI Safety & Security for Builders

A working session for solo developers, indie builders, and small teams shipping AI features who want them to be safe and reliable. I build production LLM systems and maintain an open-source AI safety library (aisafepy), so this is practical and code-first. We can add input/output guardrails, set up an evaluation harness, handle prompt-injection and jailbreak risks, deal with PII and data leakage, or review your agent/RAG setup for failure modes. Bring your repo or a short description of your stack and we'll work through it live.