Our first in-house security hire. Our advisory board gives us episodic security depth; this role brings it in-house, day to day, and helps build the automated pre-deployment tester, which is our core product.

Who

An offensive-minded security practitioner who also writes shippable code. Has actually broken LLM, agent, or web systems and wants to encode that attacker intuition into the tester, not run one-off pentests. Roughly 70% builder, 30% security leadership.

Must-haves

• 5+ years of software engineering experience, fluent in Python, with the seniority to own systems end to end.
• Red-teaming and offensive mindset, plus a solid classic application-security foundation.
• Can build, not just break: contributes production code to the attacker engine.
• LLM and agent-security literacy: prompt injection, jailbreaks, tool and MCP abuse, exfiltration via the model.
• Can turn a novel attack into a repeatable, automated test case, which is the product's actual value.

Nice-to-have

Published research or CVEs, ML internals depth, detection and defense experience, or a bug-bounty or consultancy background.