Consultant - AI Security Testing & LLM Security Assessment

KPMG Assurance and Consulting Services LLP; Navi Mumbai, Pune District, Mumbai; On-site; INR 1,200,000 - 1,800,000; Full time

Job summary

KPMG Assurance and Consulting Services LLP is seeking a Consultant to advance AI security across Generative AI, LLMs, and ML platforms. You will conduct red team exercises, assess guardrails, and review prompts to strengthen security boundaries.

The role covers evaluating AI architectures, RAG systems, and agent frameworks, with emphasis on proactive testing and secure deployment practices.

Qualifications

• Experience performing security assessments of Generative AI, LLM-based applications, AI Agents, Copilots, RAG implementations, and ML systems.
• Ability to conduct AI red team exercises and identify prompt injection, jailbreaking, data leakage, and model manipulation.
• Knowledge of guardrails, system prompts, and memory systems in AI architectures.
• Familiarity with RAG architectures, vector databases, AI Gateways, and agent frameworks.

Responsibilities

• Perform security assessments of Generative AI, LLM-based apps, AI Agents, Copilots, RAG implementations, and ML systems.
• Execute AI red team exercises to identify prompt injection, jailbreak attacks, data leakage, sensitive information disclosure, model manipulation, and AI abuse scenarios.
• Conduct adversarial testing and attack simulations against AI applications and APIs.
• Evaluate trust boundaries, attack surfaces, data flows, and model interactions.
• Assess security of LLM-based applications across OpenAI, Anthropic, Gemini, Llama, Mistral, Azure OpenAI, and other foundations.
• Review RAG architectures, Vector Databases, AI Gateways, memory systems, plugins, and agent frameworks.
• Perform guardrail effectiveness testing and validation.
• Conduct system prompt reviews and prompt hardening exercises.
• Assess Agentic AI workflows, tool calling mechanisms, orchestration layers, and autonomous decision-making controls.

Skills

Security testing

AI security

Red team

Adversarial testing

Prompt engineering

Education

CEH

OSCP

GWAPT

GPEN

CCSK

Microsoft Azure Security Certifications

CISSP

GenAI or AI Security Certifications

Prompt Engineering Certifications

Job description

Role Overview

We are seeking a highly motivated Consultant to join our AI Security practice and help organizations secure Generative AI, Large Language Models (LLMs), Agentic AI solutions, Machine Learning platforms, and AI-enabled business applications.

Key Responsibilities

AI Security Testing & Red Teaming

• Perform security assessments of Generative AI, LLM-based applications, AI Agents, Copilots, RAG implementations, and ML systems.
• Execute AI red team exercises against AI systems to identify:
• Prompt Injection
• Jailbreak attacks
• Data Leakage
• Sensitive Information Disclosure
• Model Manipulation
• Model Theft
• Training Data Extraction
• Hallucination Risks
• AI Abuse Scenarios
• Agentic AI Exploitation
• Conduct adversarial testing and attack simulations against AI applications and APIs.
• Evaluate trust boundaries, attack surfaces, data flows, and model interactions.

LLM & Agentic AI Security

• Assess security of LLM-based applications leveraging OpenAI, Anthropic, Gemini, Llama, Mistral, Azure OpenAI, and other foundation models.
• Review RAG architectures, Vector Databases, AI Gateways, MCP implementations, memory systems, plugins, and AI agent frameworks.
• Perform guardrail effectiveness testing and validation.
• Conduct system prompt reviews and prompt hardening exercises.
• Assess Agentic AI workflows, tool calling mechanisms, orchestration layers, and autonomous decision-making controls.

Desired Certifications

Preferred certifications include:

• CEH
• OSCP
• GWAPT
• GPEN
• CCSK
• Microsoft Azure Security Certifications
• ISC2 CC / CISSP
• Artificial Intelligence Security certifications
• Prompt Engineering or GenAI Certifications