AI Security Engineer

About Invoice Cloud
:

Invoice Cloud is a fast-growing fintech leader recognized with 20 major awards in 2025, including USA TODAY and Boston Globe Top Workplaces, multiple SaaS Awards wins for Best Solution for Finance and Fin Tech, and national customer service honors from Stevie and the Business Intelligence Group. Judges also highlighted our mission to reduce digital exclusion and restore simplicity and dignity to how people pay for essential services, as well as our leadership in AI maturity and responsible innovation.

It’s an award-winning, purpose-driven environment where top talent thrives. To learn more, visit

We are seeking a highly skilled and results-oriented AI Security Engineer to support the Cybersecurity, Engineering, and Data Science organizations. This role plays a critical part in advancing Invoice Cloud’s AI-first strategy by ensuring that AI/ML and generative AI systems are secure, resilient, compliant, and aligned with business objectives.

This role operates as a subject matter expert in AI security. The ideal candidate brings deep expertise in application security, AI/ML risk, and cloud-native security engineering, and serves as a trusted partner to Engineering, Product, Dev Sec Ops , Legal/Privacy, and Security Operations. Success requires strong ownership, structured problem solving, cross-functional collaboration, and the ability to balance risk reduction with business velocity.

This role is anchored in our company’s core competencies—These competencies reflect the mindsets and behaviors that define success in this role. We outline how each competency translates into real-world actions and outcomes specific to this role.

• Leads AI Security Architecture & Secure Design initiatives by designing and implementing lifecycle security controls across data ingestion, training, evaluation, deployment, and monitoring environments to measurably reduce AI‑specific risk while maintaining product velocity.
• Conducts structured Threat Modeling & Risk Assessment exercises for generative AI, RAG, and agent‑based systems, evaluating risks such as prompt injection, data poisoning, model extraction, model inversion, abuse/misuse, and data leakage, and mapping findings to OWASP Top 10 for LLM Applications, MITRE ATLAS, and NIST AI RMF to drive remediation through engineering teams.
• Defines and operationalizes Monitoring, Detection & Incident Response capabilities for AI systems by implementing prompt and output telemetry, tool‑call logging, anomaly detection, and AI‑specific incident response playbooks integrated into SIEM/SOC workflows.
• Delivers measurable outcomes aligned to 30‑, 150‑, and 210‑day milestones, including secure reference architectures, hardened AI environments, integrated security controls, and executive‑ready reporting on AI risk reduction and posture maturity.

• Establishes and formalizes AI Governance, Privacy & Third‑Party Risk requirements by defining security expectations for AI use cases, third‑party models, vendor integrations, and sensitive data usage, embedding controls into SDLC, procurement, and engineering standards.
• Drives Cross‑Functional Collaboration & Enablement by partnering with Engineering, Data Science, Dev Sec Ops , Product, Legal/Privacy, and SOC teams to align on risk appetite, escalation paths, and secure design guardrails while raising AI security maturity across the organization.
• Inventories current and planned AI/ML initiatives, documents system architectures and sensitive‑data touchpoints, and implements a structured AI security intake and risk‑rating process that ensures accountability and transparency.
• Develops and communicates forward‑looking 6‑ and 12‑month AI security maturation plans that align technical priorities with business goals and clearly articulate risk trends, metrics, and investment needs to Security leadership and the CISO.

• Integrates Secure MLOps / MLSecOps controls into AI delivery pipelines, including secure model registries, artifact signing and provenance validation, dependency scanning, secrets management, CI/CD guardrails, and hardened training and inference…