Position: Senior Security Engineer – Healthcare AI & Agent Security
Location: Bangalore
Working Days: 5 days (hybrid)
Experience Required: 5-10 years
Secure Our Client's healthcare AI agent ecosystem, ensuring safe handling of PHI, secure EHR integrations, resilient agent workflows, and enterprise-grade security for healthcare customers. Our Client's allergy and healthcare offerings include AI agents supporting clinical and operational workflows in regulated healthcare environments.
- Own the security architecture for AI agent products.
- Lead threat modeling for healthcare AI workflows.
- Review code and architecture for Python, Node.js, Java, and cloud-native services.
- Build Secure SDLC processes across engineering teams.
- Secure multi-agent systems and agent-to-agent communication.
- Assess prompt injection, indirect prompt injection, tool abuse, jailbreaks, and agent escalation risks.
- Design guardrails for PHI exposure prevention.
- Evaluate LLM, RAG, MCP, and third-party AI integrations.
- Build security testing frameworks for AI agents.
- Lead HIPAA, SOC 2, HITRUST, and customer security assessments.
- Establish PHI data handling controls.
- Define policies for healthcare data retention, access, and auditing.
- Partner with customers on security reviews and compliance questionnaires.
- Secure AWS/GCP environments.
- Secure Kubernetes, containers, APIs, and CI/CD pipelines.
- Implement CSPM and runtime security controls.
- Lead vulnerability management and incident response.
- 5–10 years in Product Security, Application Security, or Cloud Security.
- Strong experience with:
- OWASP Top 10
- API Security
- OAuth/OIDC
- Kubernetes Security
- AWS/GCP Security
- Threat Modeling
- Secure Architecture Reviews
- Hands-on penetration testing experience.
- Ability to review production code.
- Experience supporting enterprise SaaS products.
- Healthcare or HealthTech experience.
- HIPAA/HITRUST exposure.
- AI/LLM security experience.
- Experience with:
- RAG architectures
- AI agents
- MCP servers
- LangGraph/CrewAI/AutoGen-type systems
- Model evaluation and guardrails