AI Security Engineer

PwC India; Hyderabad, Gurugram District, Bengaluru; On-site; INR 2,000,000 - 5,000,000; Full time

Job summary

PwC India is seeking an AI Security Engineer to secure AI/ML pipelines, Generative AI apps and agentic AI workloads on GCP. The role blends traditional GCP security with AI-specific threats across training and inference.

Required: 5–8 years in security engineering with ~3 years in AI/ML security; hands-on with Vertex AI, GKE, Model Armor or Gemini; strong Python; TensorFlow/PyTorch familiarity; knowledge of adversarial ML; GCP Security certs preferred.

Qualifications

• Secure AI/ML pipelines against adversarial threats
• Implement security by design in training/inference
• Collaborate with ML teams to harden AI workloads

Skills

AI/ML security

Generative AI security

GCP security (Vertex AI/GKE)

Tools

Model Armor

Gemini Enterprise Security

Job description

Overview: AI Security Engineer

Secure AI/ML pipelines, Generative AI apps, and agentic AI systems on GCP; blends traditional GCP security with AI-specific threats. 5070% AI security focus.

Primary Skills

• AI/ML Pipeline Security & Adversarial ML Defense (model theft, prompt injection, data poisoning)


• Generative AI & Agentic Security (Gemini Enterprise Security, Model Armor)


• GCP AI Infrastructure Security (Vertex AI, GKE)

Secondary Skills

• Python & Secure Code Deployment


• DLP / Sensitive Data Protection (data pipeline integration)

Key Qualifications:

• Secure AI/ML pipelines against adversarial threats


• Implement Model Armor & Gemini Enterprise Security


• Secure agentic AI workflows & enterprise data access


• Harden Vertex AI & GKE for AI workloads


• Integrate DLP/SDP into training/inference pipelines


• Build Python-based security tooling


• Partner with ML teams for security-by-design


• Track OWASP LLM Top 10, MITRE ATLAS

Minimum Qualifications:

• 5–8 years in security engineering, ~3 years in AI/ML security or applied ML


• Hands-on with Vertex AI, GKE, Model Armor/Gemini


• Strong Python; familiarity with TensorFlow/PyTorch


• Understanding of adversarial ML techniques

Preferred Qualification:

• GCP certification (Security Engineer or ML Engineer) preferred