What are we building?
At
Darkhunt AI
, we're building a Cursor-like control plane for AI Security
,
a semi-autonomous system that helps teams understand how their AI agents can be attacked, why it matters, and how to defend against it.
Instead of static scans, we deploy
swarms of AI agents
that actively probe AI systems, explore attack paths, and adapt their behaviour in real time. These offensive discoveries are then translated into actionable, and increasingly automated, defensive strategies.
The complex problem is closing the loop: turning evolving attack techniques into defences that continuously learn, adapt, and harden AI systems without slowing teams down.
This is work at the intersection of
AI agents, offensive security, and defensive engineering
, where systems don't just report issues, they reason about them and respond.
This is work at the intersection of AI agents, offensive security, and defensive engineering, where systems must reason, experiment, and harden themselves.
What You'll Do?
* Build the interface where security engineers understand how AI systems are attacked and how defences evolve in response
* Design and implement
AI-native, security-critical UIs
for exploring attack chains, agent behaviours, and mitigation strategies
* Create
interaction models
for working with chats, autonomous agents, and multi-agent systems (e.g. MCP-style UI patterns)
* Turn complex, dynamic backend signals into
clear, decision-driving experiences
, not just dashboards
* Own frontend architecture decisions in a fast-moving startup, where
product, UX, and engineering are tightly coupled
Tech stack
* 16, React 19
* Material UI (MUI)
Location & Work Setup
This is a fully remote role, with the requirement to work within ±3 hours of Dublin (Ireland) time to enable close collaboration with the core team.
Who You Are
* You want to be
early, not safe
, one of the first builders shaping a product in a new and emerging AI security market
* You care about
ownership over comfort
, and you see equity as alignment with real responsibility
* You want your decisions and execution to
materially shape how the product works
, not just how it looks