Description :

Experience : 6-10 years

Location : Noida, India

Role Type : AI Security / LLM Security / Agentic Systems Security

Role Summary :

We are seeking an AI/ML Security Engineer to secure LLM-based and Agentic AI solutions. This role focuses on securing autonomous agents, mitigating prompt injection and jailbreak attacks, designing guardrails, securing RAG pipelines, and ensuring safe deployment of LLM workloads. Candidates should have strong Python skills and a good foundation in AppSec/Web Security.

Key Responsibilities :

- Identify and mitigate security risks in LLM and Agent-based systems

- Perform red teaming for jailbreaks, prompt injection, and unsafe model behaviors

- Build guardrails, content filters, and output validation layers

- Threat-model AI architectures using MITRE ATLAS & OWASP frameworks

- Secure RAG pipelines, vector databases, and embedding workflows

- Implement secure agent tool access (sandboxing, least-privilege controls)

- Collaborate with AppSec to secure APIs, deployments, and data flows

- Monitor and log anomalies and unsafe LLM interactions

- Ensure compliance with organizational and industry AI safety guidelines

Core Skills Required :

- AI/LLM Security

- OWASP LLM Top 10, Machine Learning Security Top 10.

- Prompt injection defense

- Jailbreak mitigation

- RAG + vector DB security

- Agentic system security

- Output validation & guardrails

- AI threat modeling (ATLAS)

- Security Engineering

- OWASP Top 10 basics

- Secure API design & authentication

- Vulnerability analysis & input sanitization

- Secrets management & infrastructure security basics

- Programming

- Python (advanced)

- Understanding of LangChain / AutoGen / Semantic Kernel architectures

- Familiarity with cloud AI platforms (Azure preferred)

Good to Have :

- Experience with MLOps security

- Experience with Azure OpenAI safety settings

- Containerization + sandboxing (Docker, Kubernetes)

- SOC/SIEM exposure for monitoring AI endpoints