Applied AI Security Engineer

Applied AI Security Engineer

SWIFT

London, United Kingdom

Education

University degree in Computer Science, Information Systems, or a related field; or equivalent work experience. Masters degree an asset.

Experience

Typically has 6 to 10 years of relevant work experience.

Job Responsibilities

• Run frontier AI models directly against SWIFT's codebases to find vulnerabilities that conventional tools miss
• Build the methodology for this work through direct experimentation.
• Build the pipeline that makes AI-assisted detection usable at scale: this includes writing the logic that validates findings, cuts duplicates, and routes real issues to remediation
• Run head-to-head technical evaluations of frontier models on real detection tasks, and let the evidence from testing decide which models and techniques SWIFT relies on.
• Build the integration that plugs AI-assisted detection into how SWIFT develops software, working alongside existing security tooling
• Find the vulnerability classes current tooling structurally cannot see, and build the detection approach to catch them ahead of where the rest of the industry is
• Work entirely within SWIFT's governed access and tooling controls
• Defend technical decisions under scrutiny from AI and Security peers, and document the methodology so it survives

Required Qualifications/Skills

• Demonstrated hands-on experience applying AI/LLM models to real security testing or vulnerability research
• Strong software engineering ability: can read, navigate, and reason about large, unfamiliar codebases directly, across multiple languages.
• Deep working knowledge of application security fundamentals: common vulnerability classes, exploit mechanics, and how static, dependency, and dynamic analysis tools actually work and where they fail.
• Proven ability to build working tools and pipelines from scratch, not just configure existing ones: scripting, automation, and integration work is core to this role
• Experience designing and running structured technical evaluations or experiments (model comparisons, A/B testing, or equivalent) and drawing defensible conclusions from the results.
• Comfortable operating with significant ambiguity and no established playbook; able to originate a technical approach rather than execute a known one.
• Clear technical communication: able to document and defend complex technical decisions to peer security engineers under scrutiny.

Preferred Qualifications

• Experience working in a regulated or high-assurance environment (financial services, critical infrastructure, defense, healthcare, or similarly regulated sectors).
• Familiarity with multiple frontier AI model families (e.g., GPT-class, Claude-class) and their differing strengths for security-related tasks.
• Prior experience contributing to or building a security capability that didn't previously exist, rather than operating an established one