Crack things. Professionally.

You steer the swarm on live engagements: validate the chains agents propose, push into the places autonomy can't reach — business logic, chained context, weird auth — and write the report a client can act on.

• 5+ years hands-on offensive work: red team, pentest, or adversarial research
• You've owned engagements end-to-end, from ROE to remediation retest
• You read agent output like a team lead reads a junior's work — fast, skeptical, kind

Introduce yourself →

You build the swarm: tool-using agents that enumerate, probe, and chain real attack paths. Evaluation harnesses, tradecraft libraries, guardrails, and the feedback loops that make autonomy trustworthy enough to sign.

• Strong systems engineering plus real LLM/agent experience in production
• You understand offensive tradecraft well enough to encode it — or learn fast
• You care about evaluation and evidence, not vibes

Introduce yourself →

You built this site's wormhole equivalent: WebGL, canvas, scroll choreography, and the interface where clients watch attack paths get proven live. Brand surface and product surface are one job here.

• Portfolio of shipped generative/3D interfaces you're proud of
• Fluent in the platform — no framework crutches required

Introduce yourself →

Every engagement produces a labelled corpus of real attack telemetry. You turn it into the training signal that makes the next swarm sharper — and the artifact that proves our work to clients.

• Data pipelines, labelling systems, security telemetry
• Detection engineering background strongly preferred

Introduce yourself →