Mission
The AI Security Engineer plays a key role in the InvoiceCloud Cybersecurity Program. This role requires strong attention to detail, persistence, expertise in application security and AI/ML risk, planning skills, self-motivation, organization, communication, and problem-solving abilities. The primary objective of this position is to consistently identify, prioritize, and reduce AI-specific security risks across the model lifecycle—data, training, evaluation, deployment, and operations—while maintaining business velocity and product quality.
Responsibilities
AI Security Architecture & Secure Design
Design and implement security controls for AI/ML and generative AI systems across the full lifecycle (data → training → evaluation → deployment → monitoring).
Establish secure reference architectures for common patterns (e.g., retrieval-augmented generation (RAG), model gateways, tool/agent execution) with least privilege, data minimization, and isolation.
Threat Modeling & Risk Assessment
Perform AI/ML threat modeling for new and existing systems, including prompt injection, data poisoning, model extraction, data leakage, and abuse/misuse scenarios.
Map risks to industry frameworks (e.g., OWASP Top 10 for LLM Applications, MITRE ATLAS, NIST AI RMF) and drive mitigations with engineering teams.
Secure MLOps / MLSecOps
Partner with DevSecOps/MLOps to integrate security into AI delivery pipelines (secure model registry, artifact signing, provenance, access control, dependency scanning, secrets management, CI/CD guardrails).
Ensure training and inference environments are hardened (cloud IAM, network segmentation, key management, container security).
AI Security Testing & Red Teaming
Build and execute AI security test plans and adversarial evaluations (prompt injection, jailbreaks, data exfiltration, content policy bypass, model evasion).
Develop automated test harnesses and regression suites to validate controls over time.
Monitoring, Detection & Incident Response
Define and implement telemetry for AI systems (prompt/output logging, tool calls, policy decisions) with appropriate privacy controls.
Integrate AI security signals into SIEM/SOC workflows; create detection logic and response playbooks for AI-specific incidents.
Governance, Privacy & Third-Party Risk
Support AI governance by defining security requirements for AI use cases, third-party models/vendors, and data usage.
Partner with Legal/Privacy/Compliance to ensure AI implementations align with internal policy and applicable regulations.
Cross-Functional Collaboration & Enablement
Provide security guidance, training, and documentation for engineers and data scientists; raise overall AI security maturity.
Communicate risks and progress updates to Security leadership, ELT stakeholders, and the CISO as needed.
Qualifications
This role has privileged access to highly sensitive information, intellectual property, legal matters, and complex business scenarios. The successful candidate has:
Bachelor’s degree in Computer Science, Cybersecurity, Engineering, Data Science, or related field (or equivalent practical experience).
5+ years of experience in security engineering, application/product security, cloud security, or DevSecOps.
2+ years of experience building or securing AI/ML systems (including LLM-based applications) in production environments.
Strong understanding of AI/ML threats and defenses (e.g., prompt injection, data poisoning, model extraction, model inversion, adversarial inputs, data leakage, abuse/misuse).
Experience integrating security into CI/CD and MLOps pipelines; comfortable with containerization and cloud platforms (AWS and Azure).
Preferred: Familiarity with OWASP GenAI guidance/Top 10 for LLM Applications, MITRE ATLAS, and/or NIST AI RMF.
Preferred: Certifications such as CISSP, CSSLP, CCSP, Azure Security certifications, or relevant GIAC certifications.
Personal Skills
Optimistic, persistently driving for the positive outcome
Team player; collaborative and can work independently
Excellent coordination and orchestration abilities
Strong work ethic, interpersonal skills, time management, planning and execution skills
Resourceful, collaborative, ‘out of the box’ thinking
Demonstrates a personal code of ethics, integrity, and trust
Able to successfully navigate within varying degrees of ambiguity in a fast-paced environment
Efficient communications skills (written/verbal) and interpersonal savvy
Possess a good sense of self and a strong, approachable personal presence.
Possess the determination to get results without harm, provide transparent feedback, and prioritize a positive outcome
Outcomes
First 30 days – Immersion and Formulation
Inventory current and planned AI/ML and generative AI use cases across products and internal operations; document architecture, data flows, and sensitive-data touchpoints.
Establish a baseline AI security posture by reviewing existing controls (access, secrets, logging, content filtering, data governance) and identifying immediate gaps.
Define/confirm an AI security taxonomy and intake process (risk rating, approvals, documentation, and ownership) aligned with the SDLC.
Meet with Engineering, Data Science, Product, DevSecOps, Legal/Privacy, and SOC to align priorities, risk appetite, and escalation paths.
Select and socialize the primary frameworks and references for AI security (e.g., OWASP Top 10 for LLM Applications, MITRE ATLAS, NIST AI RMF) and map them to InvoiceCloud’s environment.
Identify and prioritize the top 2–3 near-term initiatives (e.g., secure GenAI reference architecture, AI red teaming process, model and data protection controls).
Observe meetings and insert yourself into communications streams (design reviews, security reviews, SOC and incident review cadences).
First 150 days - Execution
Deliver threat models and secure design recommendations for priority AI/ML systems, including GenAI/RAG patterns and any agent/tool integrations.
Implement a secure reference architecture and guardrails for GenAI applications (prompt injection defenses, least-privilege tool access, data-loss prevention for prompts/outputs, isolation/sandboxing where needed).
Integrate AI/ML security controls into MLOps pipelines (artifact integrity, model registry protections, access control, environment hardening, dependency scanning, secrets management).
Stand up an AI security testing and red teaming workflow; build a repeatable test plan and automate regression tests for known abuse cases.
Define and implement runtime monitoring for AI systems (prompt/output/tool telemetry, abuse detection, anomaly signals) and integrate key signals into SIEM/SOC workflows.
Develop AI-specific incident response playbooks (prompt injection/data exfiltration, model theft, training data compromise, malicious output) and run at least one tabletop exercise.
Provide recurring progress updates and metrics to Security leadership and partner teams; track remediation through to closure.
First 210 days – Results
Deliver documented value for the top priorities defined in the Immersion and Execution phases (e.g., measurable risk reduction, improved guardrails, improved monitoring coverage, reduced data exposure).
Establish a sustainable operating rhythm for AI security: intake/design reviews, security testing cadence, red team findings → remediation pipeline, and monitoring/alerting ownership.
Publish a forward-looking 6‑month and 12‑month AI security maturation plan (outcomes, key metrics, tooling needs, and process improvements) aligned to business goals.
Ensure AI security requirements are embedded into product SDLC, procurement/vendor evaluations, and platform engineering standards.
Create executive-ready reporting that communicates AI security posture and trend insights to ELT stakeholders and the CISO.
Competencies
Leadership – humble but confident, persuasive, inspirational, determined, patient, accountable, and objective
Subject Matter Expert – recognized as, demonstrates, and qualifies as the as key knowledge base for risk
Business Acumen – tempers and overlays all actions, outcomes, and initiatives with a consideration of business impact, leverages knowledge of management, development, product, legal, and business development in execution of objectives
Critical, Multi-Dimensional Thinker – digs deep, doesn’t settle for surface answers or “how” answers to “what and why” questions; looks for the truth, not the easy and potentially wrong or most disruptive solution
Organization – gravitates towards bringing order to chaos and can quickly organize and maintain order from disorder
Accountability and responsibility - ability to embrace and foster a culture of ownership by accurately forecasting, reporting, and monitoring of key metrics; recognize successful achievement
Benefits
We offer a competitive benefits program including:
Medical, dental, vision, life & disability insurance
401(k) plan with company match
Flexible Time Off (FTO), wellbeing days, paid holidays, and summer Fridays
Mental health resources
Paid parental leave & Backup Care
Tuition reimbursement
Employee Resource Groups (ERGs)
InvoiceCloud is committed to providing equal employment opportunities to all employees and applicants. We do not tolerate discrimination or harassment of any kind based on race, color, religion, age, sex, nationality, disability, genetic information, veteran or military status, sexual orientation, gender identity or expression, or any other characteristic protected under applicable laws.
This commitment applies to all aspects of employment, including recruitment, hiring, placement, promotion, termination, layoff, recall, transfer, leave, compensation, and training.
If you require a disability-related or religious accommodation during the application or recruitment process, and wish to discuss possible adjustments, please contact jobs@invoicecloud.com.
Click here to review InvoiceCloud’s Job Applicant Privacy Policy.
For recruitment agencies: InvoiceCloud does not accept unsolicited resumes from agencies. Please do not forward resumes to our job aliases, employees, or any other company location. InvoiceCloud is not responsible for any fees associated with unsolicited submissions.
