Agent Security Research Engineer – Taiwan
Core
Research AI agent security risks, design detection policies, build engines/pipelines, and ship content to protect SaaS estates from agent-based attacks.
Role type
Senior IC security research engineer (AI agents)
Builds
Detection policies, engine updates, and security content for AI agents
Domain
Cybersecurity, AI Agents, SaaS Identity
Deliverable
production ML models
Required skills
Agentic tool mechanics (MCP, hooks, lifecycle), Agent-specific attack classes (prompt injection, poisoning, confused deputy), SQL (ClickHouse/Databricks/Snowflake), Python/Go scripting, SaaS/cloud identity (OAuth, tokens, scopes), Developer tooling (Git, CI/CD)
Preferred skills
Published research on LLM/agent security, Threat detection, Endpoint security (macOS/Linux/Windows), Data pipeline tooling (dbt, Dagster), SaaS security/CASB/SSPM background, Experience with Claude Code/Copilot/Cursor
Technologies
MCP, ClickHouse, Databricks, Snowflake, Python, Go, Git, GitHub, GitLab, OAuth, dbt, Dagster
Responsibilities
Research agent misbehavior and compromise vectors, design detection or policy mechanisms, build and test detection content in engines, document and help customers adopt security content, tune content based on real-world results
Seniority
Senior, hands-on IC
