微信支付-大模型安全工程师
Core
Research and develop security defenses for large language models (LLMs) and agents within WeChat Pay, focusing on prompt injection, data poisoning, and AI operational risks.
Role type
Senior LLM Security Engineer
Builds
Automated red team evaluation systems and security assessment tools for LLM applications and internal network protection.
Domain
Fintech / AI Security
Deliverable
production ML models
Required skills
LLM security, prompt injection defense, data poisoning mitigation, red teaming, security assessment tooling, internal network security, supply chain security, AI risk analysis
Preferred skills
AIGC risk understanding (hallucinations, compliance, abuse, adversarial attacks), data security, risk control strategies
Technologies
LLMs, agents, security testing frameworks, internal network security tools
Responsibilities
Design and implement defense schemes against LLM attack vectors; Build automated red team evaluation systems; Develop security assessment tools for continuous testing; Establish internal network data security systems including access control and isolation.
