Senior Security & AI Architect

SecNinjaz Technologies LLP; Delhi; On-site; INR 4,500,000 - 7,000,000; Full time

Job summary

SecNinjaz Technologies LLP seeks a Senior Security & AI Architect to blend hands-on offensive security with practical AI engineering. You will convert manual assessment methodologies into AI-assisted tools and agent workflows, validating findings with solid evidence and building secure, repeatable security capabilities.

Collaborating with customers and AI engineers, you will define scope, engagement rules, and outcomes, while refining the platform for safe, scalable security assessments.

Qualifications

• Deep hands-on offensive security experience with web apps and APIs.
• Ability to translate manual pen-testing into reusable AI tooling.
• Experience building or integrating LLM-based security workflows.
• Strong Python/backend skills for security tooling and automation.
• Knowledge of AI/LLM security risks and mitigation strategies.
• Experience collaborating with customers to define scope and outcomes.

Responsibilities

• Lead hands-on security testing, code reviews, threat modeling and architecture assessments.
• Translate assessment methodologies into reusable AI agent workflows and tools.
• Validate AI-generated findings with reproducible evidence and safety controls.
• Design secure, least-privilege workflows for AI-assisted security tasks.
• Work with customers to understand environments and security priorities.

Skills

Hands-on offensive security

AI engineering

Python

Threat modeling

Security architecture

Customer engagement

Backend programming

Tools

LLM integration

Tool calling

APIs & security tooling

Job description

Turn deep security expertise into dependable AI-powered security capabilities.

About the Role

SecNinjaz is building an AI-driven cybersecurity agent platform designed to help security teams perform assessments more efficiently, consistently, and at scale.

We are looking for a Senior Security & AI Architect whose strongest foundation is hands‑on offensive security, combined with practical experience in AI engineering, AI security, security automation, and agentic workflows.

This is not a traditional cybersecurity architect role.

The ideal candidate should understand how a highly skilled penetration tester performs assessments manually and also be capable of converting those methodologies into secure, controlled, reusable AI‑assisted tools and agent workflows.

You will work directly with customers, security practitioners, and AI engineers to understand environments, identify meaningful security risks, validate findings, and turn effective assessment techniques into product capabilities.

You will partner closely with our AI Platform Architect, who will own the underlying agent platform, orchestration, and execution infrastructure.

Your primary ownership will be:

• Offensive security methodology
• AI security and agent security
• Security validation and evidence quality
• Security automation
• AI-assisted penetration testing workflows
• Customer security assessments
• Translating security expertise into reusable engineering capabilities

What You Will Own

• Lead hands‑on Web Application and API penetration testing, source‑code security reviews, threat modeling, and security architecture assessments.
• Identify and validate realistic attack paths involving authentication, authorization, access control, injection, business logic, APIs, application architecture, and sensitive‑data exposure.
• Work directly with customers to understand their environment, security priorities, assessment scope, rules of engagement, technical constraints, and expected outcomes.
• Convert manual penetration‑testing methodologies into reusable security tools, automated test cases, and bounded AI agent workflows.
• Define clear preconditions, permissions, stopping conditions, safety controls, evidence requirements, and human‑approval checkpoints for security agents.
• Personally build and integrate AI‑assisted security capabilities using LLM/model APIs, tool calling, APIs, structured outputs, security tools, and deterministic validation.
• Design workflows where AI agents can assist with reconnaissance, vulnerability analysis, code review, attack‑path analysis, evidence collection, triage, remediation guidance, and remediation validation.
• Independently validate AI‑generated security findings using reproducible technical evidence rather than trusting model‑generated conclusions.
• Define how the platform distinguishes between:

• Hypotheses
• Validated vulnerabilities
• False positives

• Develop representative security test cases, evaluation datasets, reproducible validators, and regression tests for AI‑driven security capabilities.
• Test AI and agent trust boundaries, including prompt injection, indirect prompt injection, unsafe tool execution, excessive permissions, credential exposure, sensitive‑data leakage, malicious inputs, and unintended agent actions.
• Help design and validate controls around least privilege, tool authorization, credential management, isolation, sandboxing, auditability, and human approval.
• Help design secure security knowledge and assessment memory with appropriate provenance, customer isolation, access control, freshness, validation, and lifecycle management.
• Support customer pilots, security integrations, assessment delivery, troubleshooting, and implementation.
• Convert lessons learned from real customer environments and security assessments into reusable product improvements.
• Contribute production‑quality code, review implementations, mentor engineers, and communicate security findings clearly to both technical teams and decision‑makers.

Must‑Have Qualifications

• Deep hands‑on offensive security experience, including Web Application and API penetration testing, source‑code review, realistic attack‑chain development, and remediation verification.
• Strong practical expertise in authentication, authorization, injection vulnerabilities, business‑logic vulnerabilities, access‑control weaknesses, API security, and application security architecture.
• Experience translating manual penetration‑testing methodologies into reusable security tooling, automation, test cases, or controlled AI agent workflows.
• Hands‑on experience building or integrating LLM/AI agents with security tools, APIs, external services, structured outputs, or automated cybersecurity workflows.
• Strong understanding of AI/LLM security risks, including:

• Unsafe tool execution
• Credential exposure
• Sensitive‑data leakage
• Untrusted inputs
• Agent manipulation

• Ability to independently validate AI‑generated security findings using reproducible technical evidence rather than relying solely on model responses.
• Strong Python and/or backend programming skills, with demonstrated experience building security tooling, automation, integrations, or production‑quality security software.
• Experience with threat modeling and secure architecture, including least privilege, credential protection, isolation, secure tool execution, authorization, auditability, and human‑approval controls.
• Practical AI engineering experience. You should have personally built a tool‑using LLM workflow, AI agent, AI‑enabled security capability, or similar system and be able to explain its architecture, integration, evaluation, limitations, and failure modes.
• Experience with backend APIs, databases/data stores, authentication, credentials, deployment, logging, and integrations sufficient to build and troubleshoot secure AI‑security systems.
• Experience working directly with customers or stakeholders to define assessment scope, rules of engagement, security priorities, evidence requirements, remediation plans, and validation criteria.
• Strong senior‑level technical judgment with the ability to review architecture, review code, prioritize security risks, mentor engineers, and translate security expertise into maintainable engineering capabilities.

Certification Requirement

• OSCP or an equivalent/higher‑level verifiable hands‑on offensive‑security certification is required.
• Certification must be supported by demonstrated practical penetration‑testing and security‑engineering capability.
• Certification alone will not be considered sufficient without strong hands‑on technical experience.

Preferred Qualifications

• OSWE, OSEP, OSED, OSCE3, CREST CRT/CCT, or other advanced hands‑on offensive‑security certifications.
• Experience performing AI Red Teaming or security testing of LLM applications, AI agents, and agentic systems.
• Experience securing or testing:

• LLM applications
• RAG systems
• AI agents
• Multi‑agent systems
• Tool‑calling architectures
• MCP integrations
• AI memory systems
• AI security evaluation frameworks

• Experience with advanced white‑box application security testing.
• Experience with exploit development, vulnerability research, reverse engineering, or advanced web exploitation.
• Experience building cybersecurity agents or automation for:

• VAPT
• Vulnerability analysis
• Source‑code review
• Attack‑path analysis
• Security triage
• Evidence collection
• Remediation validation
• Security reporting

• Experience with cloud security, identity security, infrastructure security, container security, or multi‑tenant isolation.
• Experience building reusable security automation or vulnerability‑management workflows.
• Published CVEs, security research, bug‑bounty achievements, open‑source security tools, technical blogs, conference talks, or other meaningful security contributions are an advantage.

We are specifically looking for candidates who can combine deep offensive‑security expertise with practical AI engineering.

Using ChatGPT, coding assistants, or security tools alone does not demonstrate AI engineering experience.

The successful candidate should understand both how a skilled security practitioner performs an assessment and how selected parts of that expertise can be safely transformed into scalable AI‑assisted product capabilities.