Senior AI Security Engineer

About bunny.net Some people accept the internet as it is. We see what it could be. At bunny.net, we don’t wait for change. We make it happen. You won’t find layers of red tape or play-it-safe thinking here. You’ll find people who move quickly, take ownership, and push ideas all the way to the edge. Literally. You’ll build things that solve actual problems, not add to them. We’re not perfect, and we don’t pretend to be. But we’re always improving. We’re looking for people who want to do the same, who take pride in their work, and aren’t afraid to challenge what’s possible. Because everyone deserves a better internet. And we’re here to help it hop faster. Senior AI Security Engineer AI agents are changing how software gets built and operated, and they’re opening a whole new attack surface. Security is racing to catch up. We’re building a security layer for that traffic at the edge. It will inspect requests going into LLMs and the responses coming back, detect threats in real time, and block them without adding noticeable latency. This is a new area for us, and there’s still a lot to figure out. You’ll work with a small AI team to define what we detect, which techniques work in production, how we keep false positives under control, and how we do all of that across bunny.net’s global edge network. What will you do? Shape AI security at the edge , turning new threats into protections customers can actually use Think like an attacker , finding new ways AI systems can be exploited, and ways to stop them Build for real time , catching threats without slowing down the applications we’re protecting Own problems end-to-end , from early experiments and architecture to production Make the hard trade-offs between detection accuracy, false positives, latency, cost, and customer experience Take what works globally across bunny.net’s edge network What will you bring? 7+ years building and architecting production software , with hands-on experience in Python Deep hands-on LLM and agent experience , including model APIs, tool/function calling, context management, guardrails, and evaluation Strong LLM security experience , including detecting threats such as prompt injection, jailbreaks, data leakage, tool abuse, or agent security Hands-on machine learning experience , building or integrating models for classification, anomaly detection, or threat detection Strong HTTP and API knowledge , ideally from building gateways, proxies, middleware, or systems that operate in the request/response path You use AI tools to move faster, and you know how to test, challenge, and safely use what they generate in production What will make us extra happy? Real-time detection or enforcement experience , working with rules, heuristics, classifiers, anomaly detection, ML models, or a combination of them Production experience with LangChain, LangGraph, Hugging Face, PyTorch, TensorFlow, scikit-learn, or similar Experience with model…