AI Security Engineer: Safeguard AI-Driven Apps

Academic Key; United Kingdom; Remote; GBP 70,000 - 110,000; Full time

Job summary

BJAK is expanding its AI security program in the United Kingdom, seeking an AI Security Engineer to protect AI-enabled products and agent workflows. You will assess data shared with model vendors, implement least-privilege controls, and test for prompt injection across prompts, history, and untrusted content.

You will design tenant and user data isolation, collaborate with Product and Engineering on safe ingestion and output validation, and support regulatory controls across SC TRM and BNM RMiT.

Qualifications

• Degree in Computer Science, Cybersecurity, AI, or related field or equivalent experience.
• 3+ years in application security, product security, security engineering, or AI system security.
• Experience implementing SC TRM and BNM RMiT controls or equivalent regulatory evidence.
• Understanding of third-party LLM integrations and AI architectures.
• Programming or scripting skills, preferably Python and TS/Node.js, plus APIs and cloud platforms.

Responsibilities

• Secure AI-enabled products and agent workflows with least-privilege controls and explicit authorization.
• Assess data shared with model vendors, minimize exposure, and ensure data isolation.
• Test for prompt injection and indirect injection across prompts, history, and documents.
• Collaborate with Product and Engineering on safe data ingestion, output validation, and approvals.
• Support risk management and regulatory evidence with practical controls and communications.

Skills

Python

TypeScript/Node.js

APIs

AWS

GCP

English communication

Education

Degree in Computer Science, Cybersecurity, AI, or related field

Tools

LLM Integrations

Cloud APIs (AWS/GCP)

Job description

BJAK is expanding its AI security program in the United Kingdom, seeking an AI Security Engineer to protect AI-enabled products and agent workflows. You will assess data shared with model vendors, implement least-privilege controls, and test for prompt injection across prompts, history, and untrusted content.

You will design tenant and user data isolation, collaborate with Product and Engineering on safe ingestion and output validation, and support regulatory controls across SC TRM and BNM RMiT.