Teams at Singapore AI Safety Hub

Develop and productionize zero-knowledge verification systems for frontier AI inference. Responsibilities include reducing proving costs, building proof circuits, emulating GPU floating-point behavior, designing sampling schemes, red-teaming constraint systems, validating soundness, publishing research, and collaborating with cryptographers, hardware-security specialists, policymakers, and systems engineers.

Design and evaluate software-only protocols for verifying AI datacenter compute and memory usage. The role involves developing challenge kernels, microbenchmarking GPUs, measuring latency across cluster topologies, studying compression and hidden capacity, red-teaming timing assumptions, and publishing findings. Work spans GPU performance engineering, adversarial measurement, cryptographic protocol design, hardware security, and LLM inference systems.

Own and scale the operations function for a rapidly growing international AI safety organization. Responsibilities include upgrading hiring, payroll, onboarding, expenses, office and applicant-management systems; establishing performance, talent, project-management and benefits processes; managing UK and Singapore entity infrastructure, compliance, offices and security; and partnering with leadership on culture, strategy, governance and policies. The role requires proactive execution, people leadership, stakeholder management, regulatory judgment and organizational scaling experience.