Member of Technical Staff, AI Products

You will build the model side of our platform: the agents, retrieval, and reasoning pipelines that take raw security data and produce a verdict an analyst can act on in seconds.

This is not a research role. You will own end-to-end shipping. From prototype on a real customer dataset, to evals, to a feature in production behind real traffic. Weekly cycles, not quarterly.

• Design and ship agent systems that triage, investigate, and respond to security alerts
• Build evals on real customer data so we know what good looks like before we ship
• Work side by side with detection engineers to ground the models in real attacker behavior
• Own the AI quality bar end to end: latency, cost, accuracy, hallucination, and recovery

• Four or more years building production software, with recent applied AI work
• Comfortable owning a feature from spec to launch with no hand-holding
• Strong instinct for evals, regression tests, and shipping things that work next week
• You read papers when they help and ignore them when they do not

• Past work on LLM agents, retrieval, or tool-using systems
• Background in security, SOC operations, or detection engineering
• Open-source contributions in the AI or security space