Teams at Simular

Own end-to-end security for AI computer-use agents, including application security, cloud and Kubernetes hardening, CI/CD security, sandboxing, prompt-injection defenses, incident response, vendor reviews, and compliance. Build security tooling, establish foundational processes, partner with engineering and research, and protect the evolving attack surface of autonomous AI agents.

Build the low-level control layer that translates AI intent into reliable computer actions across macOS, Windows, and Linux. Develop mouse and keyboard injection, screen capture, UI element detection, accessibility integrations, cross-platform abstractions, and error recovery. Address targeting, focus, scaling, occlusion, and timing issues while collaborating with model and planning teams on intent execution interfaces.

Build and maintain dashboards tracking activation, retention, task success, and agent reliability; analyze agent run data to locate failures and stalls; design instrumentation and metrics; support experiment design and readouts; produce reproducible analyses and actionable recommendations for product and engineering.