Principle AI Architect

Join Trend ‧ Join New Generation

趨勢科技 - 全球雲端資安領航者 / 全亞洲最大軟體公司 / 企業版圖橫跨五大洲 / 趨勢全球研發基地在台灣

===============================================================

About This Role

Trend Micro’s security products protect hundreds of millions of devices and cloud workloads worldwide. We have accumulated decades of threat intelligence - across endpoint protection, vulnerability research, and large-scale XDR telemetry - and we are now building the AI-native systems that encode, amplify, and operationalize that expertise at scale.

We are looking for a Principal AI Architect who brings deep security domain instinct, production AI/ML experience, and the external credibility to represent Trend Micro’s technical thinking at the world’s most demanding security venues.

This is not a delivery role. You are the technical conscience of AI integration across the organization - reviewing proposals, shaping architecture, distilling security expertise into AI-legible systems, and ensuring that AI adoption genuinely improves how engineers and products work. You will be embedded in one of three domain tracks (see the end of this document), each with distinct technical depth and external focus.

Core Responsibilities

AI Architecture

Own the architectural vision for how AI/ML integrates into your domain’s product stack - on-device inference, cloud-assisted pipelines, and agentic decision workflows.

• Define tradeoff frameworks for AI capabilities within real production constraints: resource budgets, latency, evasion resistance, and platform-specific requirements
• Assess and review AI feature proposals for architectural soundness, data dependency risks, and long-term maintainability
• Identify where consolidation or modernization initiatives create AI architecture opportunities - and where they introduce risk
• Design the boundary between automated AI decision-making and human expert judgment in critical workflows

Knowledge Distillation

Work directly with senior engineers, threat researchers, and domain experts to extract tacit expertise and encode it into AI-legible form.

• Translate behavioral heuristics, detection intuition, and domain-specific patterns into model training signals, evaluation suites, and agentic decision frameworks
• Build evaluation frameworks that test whether AI systems actually replicate expert judgment - not just surface-level pattern matching
• Prevent knowledge loss as senior experts transition - systematically encoding expertise that currently retires with the person

AI Adoption Quality

Define what meaningful AI adoption looks like for engineering and research teams - beyond token usage, into workflow integration and architectural thinking.

• Build evaluation frameworks that engineering leadership can act on
• Identify engineers and teams who are genuinely advancing AI-native development and help systematize what they are doing
• Advise on AI tooling that genuinely accelerates work, and flag where AI introduces risk rather than leverage

External Representation

Represent Trend Micro R&D at the industry venues most relevant to your domain - GTC, Black Hat, DEF CON, RSA, AWS re:Inforce, Pwn2Own, and academic/standards forums.

• Contribute to external discourse through technical blog posts, conference presentations, and original research
• Engage with MITRE ATLAS, OWASP frameworks, and relevant standards bodies to shape how AI security practices are defined
• Operate as a technical peer to CTO-level counterparts at partners, customers, and the Anthropic enterprise relationship

What You Bring

Security Domain Depth

• Hands-on background in a security engineering or research discipline - endpoint platforms, vulnerability research, threat intelligence, or cloud/XDR architecture
• Deep understanding of the internals of at least one major security domain at production scale
• Has shipped or contributed to real security systems with real-world constraints

AI / Agentic Fluency

• Production experience with SLM/LLM integration, agentic workflow design, or multi-model systems
• Can reason about inference tradeoffs - latency, privacy, accuracy, evasion resistance - in context
• Understands the full AI pipeline: data, model, integration, deployment, monitoring

Independent Judgment

• Pushes back when architecture doesn’t hold up - not swayed by schedule pressure or demo quality
• Represents engineering truth, not the convenient answer
• Comfortable being the neutral technical voice in rooms where everyone else has a delivery stake

External Credibility

• Has spoken at, published in, or contributed to recognized security or AI venues
• Recognized - or on a clear trajectory to be recognized - as a technical voice in the field

What This Role Is Not

• Not a team lead or delivery owner
• Not a hands-on daily coder of production systems
• Not a pure ML researcher - training models from scratch is not the focus
• Not a sales or pre-sales function

Nice to Have

• Experience with adversarial ML, model integrity, or AI system threat modeling
• Familiarity with AI-native security product development
• Prior work at or collaboration with TrendLabs, ZDI, or equivalent threat research organizations
• Published CVEs, conference papers, or open source security tooling
• Background spanning both offensive security and AI/ML systems

Domain Tracks

This role has three domain tracks. Candidates will be placed in one based on background and organizational fit. All tracks share the same seniority, scope, and reporting structure - the difference is the security domain and the external community each person represents.

Endpoint Security - AI architecture for endpoint agent consolidation, on-device inference, and Windows/Linux platform integration. External focus: GTC, Black Hat, DEF CON, Pwn2Own.

Threat Intelligence - Knowledge distillation from TrendLabs and ZDI researchers, AI-assisted vulnerability analysis, and threat research × AI architecture. External focus: Pwn2Own, Black Hat, DEF CON, academic security research.

Cloud / XDR - AI integration across XDR correlation, CNAPP/CSPM pipelines, and large-scale security telemetry. External focus: AWS re:Inforce, Google Next Security, RSA, GTC.

The hiring process will clarify which track is the best fit. Candidates with cross-domain experience are encouraged to apply.

===============================================================

連結智慧 守護世界 --- Connected Intelligence for Securing a Connected World