We are looking for a Senior Staff AI Engineer to take a technical leadership role in designing, evolving, and owning OKX's core security AI capabilities. This is an end-to-end, research-to-production role: you will set technical direction, own algorithms across multiple high-impact domains, and ship models into real-money, high-risk, large-scale production environments
You will operate as both a hands-on technical contributor and an applied-science leader — driving complex cross-team initiatives, mentoring senior engineers, and ensuring our AI solutions are robust, efficient, and aligned with company strategy. Your domains span
Face intelligence & liveness detection — defenses against presentation attacks, injection attacks, and DeepFakes
Image & video quality assessment — reliable, security-critical decisions across diverse devices, environments, and network conditions
Document understanding & OCR — identity document parsing and FakeID detection for verification and compliance use cases
Multimodal risk & identity intelligence — fusing visual, behavioral, and contextual signals into identity-quality and trust signals
AI agents — applying agentic AI to scale investigation, detection, and operational defense
Lead strategic AI initiatives: Drive the end-to-end design, training, and deployment of state-of-the-art deep-learning and multimodal models that scale to millions of users — spanning computer vision, vision-LLM, multimodal fusion, and AI-agent applications
Architect scalable models & systems: Build and own production-grade AI pipelines for OCR / text extraction, biometric matching, fraud-pattern recognition, and AIGC generation/detection — engineered for low latency, high availability, and strong reliability
Own face intelligence & liveness: Lead anti-spoofing capabilities against presentation, injection, and DeepFake attacks, and develop image/video quality models that keep security decisions reliable in real-world conditions
Build identity & trust signals: Leverage user behavior, historical patterns, and cross-modal risk indicators to produce robust identity-quality and trust signals
Drive execution & collaboration: Deliver high-quality, measurable results while managing ambiguity and prioritizing technical debt; partner with backend, product, infrastructure, risk, and compliance teams to integrate AI into large-scale systems, and influence stakeholders across technical and non-technical backgrounds
Advance applied R&D: Drive experimentation in vision-LLM, multimodal representation learning, and AI agents; stay current with academic and industry advances and translate them into practical, production-ready defenses
Elevate the technical bar & mentor: Provide technical guidance and mentorship, set a high engineering standard, and democratize applied-AI best practices across the team and the wider organization