Agent Security Engineer

Job overview

The Agent Security Engineer role is listed in Brazil, IN, and is described as onsite. Compensation is not specified. Belmont Lavan is hiring to secure autonomous AI agents before and after they reach production. Positioned between AI engineering and security, this engineer designs guardrails, tests agent systems through offensive security work, and builds governance controls. The work helps the business deploy agents with confidence by addressing identity, access, monitoring, and AI-specific attack risks.

What you'll do

• Design security guardrails and governance controls for autonomous AI agents
• test agent systems and address risks including prompt injection, jailbreaks, tool-based data exfiltration, and RAG poisoning
• design identity and access controls
• apply cloud security, logging, detection, and policy-as-code practices
• communicate AI risk to engineers, risk teams, and executives.

What we're looking for

Skills & competencies

agent security engineer

brazil in

belmont lavan

ai agent security

llm security

application security

cloud security

identity and access management

oauth 2.0

oidc

python

langgraph

openai agents sdk

model context protocol

Work arrangement

• Weekend coverage required

Why this role

Requires 5+ years in security engineering, including at least 1 year securing LLM or agent-based systems; hands-on knowledge of LLM attack techniques; Python and one of Go, TypeScript, or Rust.

About the employer

Belmont Lavan Ltd is hiring for this role. Industry: Other Scientific and Technical Consulting Services. Sector: 54.