Associate Director AI Security Engineering

Job Title: Associate Director AI Security Engineering

Location: Hyderabad or Mumba

iJob ID: B-26034

EC-Council is the world’s largest cyber security technical certification body. We operate in 145 countries globally and we are the owner and developer of various world-famous cyber security programs. We are proud to have trained and certified over 400,000 information security professionals globally that have influenced the cyber security mindset of countless organizations worldwide.

www.eccouncil.org

Role Overview:

We are looking for an Associate Director – AI Security Engineering to lead the design, implementation, and governance of security controls across our AI/ML and agentic systems. This role sits at the intersection of AI engineering, cybersecurity, software engineering, and governance — responsible for securing everything from LLM applications and RAG pipelines to autonomous AI agents, fine-tuned models, and the infrastructure that serves them.You will build and lead a team responsible for identifying, testing, and mitigating security risks unique to AI systems, while also embedding security-by-design practices into the broader AI development lifecycle.

Key Responsibilities:

Strategy & Leadership

• Define and drive the AI security engineering roadmap, standards, and governance frameworks across the organization
• Build, mentor, and lead a team of AI security engineers and researchers
• Partner with AI/ML engineering, platform, product, and compliance teams to embed security throughout the AI development lifecycle
• Own AI risk assessments, threat modeling, and security sign-off for new AI/agentic system launches

AI Application & LLM Security

• Lead defenses against prompt injection, jailbreaks, role overrides, system prompt extraction, prompt leakage, multi-turn manipulation, and context poisoning
• Oversee security architecture for RAG pipelines — protecting against retrieval poisoning, malicious documents, hidden prompt payloads (HTML/Markdown injection), vector/embedding poisoning, chunk manipulation, and cross-document contamination
• Establish secure prompt engineering and evaluation standards across product teams

Agentic AI & Tooling Security

• Design and enforce security models for AI agents, including tool permissioning, least privilege, identity/authentication, session isolation, human-in-the-loop approval workflows, multi-agent trust boundaries, and memory isolation
• Lead Model Context Protocol (MCP) security initiatives: server trust validation, authorization boundaries, capability restrictions, context/input sanitization, and secure tool sandboxing
• Drive Security Testing for Agentic AI Systems — including SAST, DAST, and SCA adapted to AI/agentic architectures — across the CI/CD pipeline

Model & Data Security

• Secure fine-tuning pipelines (SFT, LoRA, QLoRA, PEFT) against dataset/label poisoning, backdoors, and malicious training data
• Mitigate model-level risks: model theft, model extraction, membership inference, model inversion, and weight tampering
• Define secure practices for model serving platforms (vLLM, Ollama, TGI) including access control and runtime hardening

Vector Database & Infrastructure Security

• Establish security standards for vector databases (Qdrant, Pinecone, Weaviate, Milvus, Chroma) — encryption, tenant/namespace isolation, metadata filtering, access control, index poisoning prevention, and secure deletion/retention policies
• Secure the broader AI attack surface: APIs, agent memory, browsers, code execution sandboxes, databases, and third-party/external service integrations

Governance & Compliance

• Implement auditability, explainability, and policy enforcement mechanisms across AI systems
• Support regulatory and internal audit requirements related to AI risk, data protection, and responsible AI use
• Produce security assessments, incident reports, and executive-level risk communications

Required Qualification

• 10+ years in security engineering, with 3+ years in a technical leadership or management role
• Demonstrated hands-on experience securing LLM-based applications, RAG systems, and/or AI agents in production
• Strong understanding of transformer architecture, attention mechanisms, tokens, context windows, embeddings, and semantic search
• Practical experience with fine-tuning methods (SFT, LoRA, QLoRA, PEFT) and their associated security risks
• Familiarity with model serving fram…