AI Security Solution Architect - Guardrail Engineering & Adversarial Validation

About Cognizant Corporate
Cognizant Corporate is a global community united by a shared purpose\: to make a meaningful impact. We are committed to excellence and driven by outcomes that matter. Collaboration is at the heart of how we work, and our forward-thinking mindset fuels continuous learning, innovation, and growth.

At Cognizant, careers transcend titles. We empower our people to think strategically, inspire others, and lead with purpose, always guided by our core values. Join us in shaping future of business. Could you be the one to make a difference?

About the role
As an AI Security Solution Architect - Guardrail Engineering & Adversarial Validation, you will drive impactful contributions and focus on outcomes. You will be a key member of the AI Security Engineering & Enablement (AISE) team, collaborating with platform security architects, hyperscaler architects, delivery teams, DevSecOps teams, AI workload owners, and security stakeholders. You will have the autonomy to lead, innovate, and improve workflows while upholding our commitment to quality and continuous learning. As a team of self-starters, you can work with impact with our vibrant people and culture while enjoying unmatched learning opportunities.

In this role, you will\:

• Design, build, and maintain AISE’s guardrail library, including reusable prompt sanitization SDKs, output filtering functions, PII/PHI redaction modules, and topic boundary enforcement components.
• Develop and maintain AISE’s adversarial test automation framework covering OWASP LLM Top 10 and MITRE ATLAS TTPs, including prompt injection, indirect prompt injection, RAG exfiltration, jailbreak, model extraction, agent abuse, and supply chain attacks.
• Operate AISE’s adversarial validation tollgate by running adversarial tests against AI workloads submitted for production review, documenting findings and severity, issuing residual risk statements, and supporting production approval decisions.
• Integrate AI security controls into DevSecOps pipelines by building pipeline stages that enforce prompt validation, dependency scanning, model integrity checks, and AI workload security requirements.
• Engineer reusable AI security capabilities, including IaC security modules, SIEM telemetry, anomaly signatures, detection logic, and implementation-ready recommendations for AI development pipelines.

Embrace our vibrant culture by striving for excellence, focusing on meaningful outcomes, and collaborating effectively. Prioritize quality, stay curious, and commit to continuous learning. Take ownership, build relationships, and focus on personal growth to drive business strategy, creating unmatched career opportunities and impactful work.

What you must have to be considered

• 10+ years of experience in security engineering or application security, including 3+ years directly in AI/ML security, red team, or AI security tool development.
• Hands-on development experience with the ability to write and ship production Python, with TypeScript or Go experience preferred.
• Direct experience integrating or extending at least two AI security tools such as Garak, PyRIT, LLM Guard, Presidio, NeMo Guardrails, Lakera Guard, or equivalent tools.
• Practical experience executing prompt injection, indirect prompt injection, and jailbreak attacks against real LLM applications in red team, research, or CTF environments.
• Experience with RAG pipeline security, including retrieval-based data exfiltration, embedding inversion, vector store poisoning, and related defense patterns.
• Experience with agentic system adversarial testing, including agent loop hijacking, tool abuse, and privilege escalation in multi-agent orchestration frameworks.
• Solid understanding of OWASP LLM Top 10 and MITRE ATLAS, with the ability to map TTPs to concrete adversarial test cases.
• Experience integrating security into CI/CD pipelines such as GitHub Actions, Azure DevOps, or Jenkins by building security pipeline stages.
• A strong sense of ownership, desire to create meaningful outcomes, and passion for work that serves a greater good for customers, communities, or global challenges.

These will help you succeed

• Experience developing SIEM detection logic for AI workloads, including Sigma rules, KQL, or SPL queries for AI-specific anomaly detection.
• IaC security experience, including writing or reviewing Terraform or Bicep modules with security controls encoded.
• Familiarity with LangChain, LangGraph, LlamaIndex, or Semantic Kernel at the code level, including how these frameworks handle tool calls, memory, and retrieval.
• Published research, CVE disclosures, blog posts, conference talks, or other externally recognized contributions in AI or LLM security.
• Experience in a security engineering or red team function within a large enterprise or consulting organization.
• Ability to communicate adversarial findings to senior security leaders while partnering directly with development teams to implement mitigations.

Work model - Hybrid
We strive to provide flexibility wherever possible. Based on this role’s business requirements, this position is hybrid with 2-3 days work from office. Regardless of your working arrangement, we are here to support a healthy work-life balance through our various wellbeing programs.

The working arrangements for this role are accurate as of the date of posting. This may change based on the project you’re engaged in, as well as business and client requirements. Rest assured, we will always be clear about role expectations.

We're excited to meet people who share our mission and can make an impact in a variety of ways. Don't hesitate to apply, even if you only meet the minimum requirements listed. Think about your transferable experiences and unique skills that make you stand out as someone who can bring new and exciting things to this role