Lead AI Security Engineer-APS - IT (Information Technology) - Bangalore, India
Societe Generale · Bangalore, India
About this role
We are seeking a Lead AI Security Engineer to drive the security of next-generation AI systems, GenAI platforms, LLM-powered applications, and AI-enabled enterprise solutions. The role combines Artificial Intelligence, secure software engineering, and offensive application security to identify, assess, and mitigate risks across AI applications and modern software ecosystems. This role is focused on securing AI-led applications and platforms rather than performing traditional infrastructure penetration testing.
Responsibilities
Lead AI security assessments for GenAI applications, LLM integrations, AI agents, RAG architectures, and AI-powered enterprise platforms.
Perform AI red teaming to identify prompt injection, jailbreaks, indirect prompt attacks, data leakage, insecure tool use, agent abuse, RAG poisoning, and model interaction risks.
Design adversarial test cases and evaluation methods to validate AI system resilience, trust boundaries, and guardrail effectiveness.
Define practical mitigation patterns for AI risks, including secure prompt design, policy controls, input/output validation, grounding checks, logging, monitoring, and abuse prevention.
Build AI-powered security automation for vulnerability discovery, source code analysis, threat analysis, test generation, and remediation acceleration.
Integrate AI security validations into SDLC, DevSecOps, CI/CD, SAST, DAST, and code review pipelines.
Conduct secure design reviews, threat modeling, code reviews, and vulnerability assessments for AI-led applications, APIs, microservices, and cloud-native platforms.
Lead AI security initiatives, define standards, and mentor engineers on AI security and secure coding practices.
Required qualifications
8+ years of experience across software engineering, AI engineering, cybersecurity engineering, or application security.
Strong hands-on experience with GenAI, LLMs, AI agents, RAG architectures, embeddings, vector databases, and AI application patterns.
Development experience in Python, Java, Go, JavaScript/TypeScript, C#, or similar languages.
Practical knowledge of penetration testing, SAST, DAST, secure code review, threat modeling, and vulnerability validation.
Experience securing AI workloads or applications on Azure, AWS, or GCP.
Knowledge of Kubernetes, OpenShift, Docker/container security, secrets management, and identity security.
Preferred qualifications
Experience securing GenAI platforms, AI agents, copilots, LLM applications, or AI-driven cybersecurity products.
Experience building AI-assisted security tooling, AI-based code analysis systems, or automated security testing platforms.
Certifications such as AI security/red teaming certifications, OSCP, OSWE, GWAPT, CISSP, CSSLP, GIAC, or equivalent practical experience.
Background in vulnerability research, AI red teaming, application security research, or secure AI architecture.
For full details about this role, click “Apply on company site” to view the employer’s original posting.
A focused preparation guide, built around this job’s responsibilities and requirements.
✦ AI-generated guide
Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.
01
Understand the role
What matters most
AI Red Teaming and Adversarial Testing
The role requires leading assessments to identify vulnerabilities specific to GenAI, such as prompt injection, jailbreaks, and RAG poisoning. You must demonstrate the ability to design adversarial test cases that validate the resilience of AI models and guardrail effectiveness beyond traditional application security.
AI-Powered Security Automation
A core priority is building custom security tooling and automation frameworks. You will be expected to develop systems for automated vulnerability discovery, code analysis, and remediation acceleration, integrating these directly into CI/CD and DevSecOps pipelines.
Secure AI Architecture and Governance
You must define and implement practical mitigation patterns for AI risks, including secure prompt design, input/output validation, and logging. This involves partnering with engineering teams to embed secure-by-design principles into AI-led applications and cloud-native platforms.
Likely challenges to think through
Balancing Innovation with Security Guardrails
AI development moves rapidly, and implementing strict security controls can often be perceived as a bottleneck. Prepare to discuss how you balance the need for rigorous security validation and guardrails without stifling the velocity of AI engineering teams.
Adapting Traditional AppSec to Non-Deterministic Systems
Unlike traditional software, AI models are non-deterministic, making standard testing methodologies less effective. You will face the challenge of adapting your deep knowledge of OWASP and traditional AppSec to the unique, probabilistic nature of LLMs and AI agents.
03
Practice your answers
Technical practice
Mitigating Prompt Injection in RAG Architectures
What to coverExplain how you would design input validation and output filtering to prevent prompt injection in a Retrieval-Augmented Generation (RAG) system. Discuss the trade-offs between strict input sanitization and maintaining the utility of the LLM's natural language processing capabilities.
AI Red Teaming Methodology
What to coverDescribe your approach to performing AI red teaming for an LLM-powered application. Focus on how you identify risks like jailbreaks, indirect prompt attacks, and insecure tool use, and how you document these findings for engineering teams.
Integrating AI Security into DevSecOps
What to coverHow would you integrate automated AI security validations into a CI/CD pipeline? Suggest specific tools or frameworks for SAST/DAST that can be adapted for AI-specific vulnerabilities like model poisoning or data leakage.
Securing AI Agents and Tool Use
What to coverWhen an AI agent has access to external tools or APIs, what security controls should be implemented to prevent unauthorized actions? Discuss the concept of least privilege in the context of LLM tool execution.
Adversarial Testing and Guardrails
What to coverHow do you design adversarial test cases to validate the effectiveness of guardrails in a production AI environment? Explain how you measure the resilience of these guardrails against evolving attack vectors.
Cloud-Native Security for AI Workloads
What to coverGiven the requirement for Kubernetes and container security, how do you secure the infrastructure hosting AI models and vector databases? Address secrets management and identity security for microservices interacting with AI APIs.
AI-Assisted Security Automation
What to coverDescribe a scenario where you would use AI to accelerate vulnerability discovery or code analysis. How do you ensure the reliability and accuracy of the security insights generated by an AI-based tool?
Threat Modeling for AI-Enabled Platforms
What to coverHow does threat modeling for an AI-led application differ from traditional web application threat modeling? Highlight the specific components of an AI system, such as the retrieval layer or model interaction, that require unique focus.
Behavioural practice
Leading Cross-Functional Security Initiatives
What to coverDescribe a time you led a security initiative that required collaboration between engineering, architecture, and cybersecurity stakeholders. How did you align these groups on security standards for a complex technical project?
Mentoring Engineers on Secure Coding
What to coverProvide an example of how you have mentored engineers on secure coding practices. How did you adapt your communication style to ensure the technical guidance was actionable and well-received?
Handling Disagreements on Risk Prioritization
What to coverTell us about a time you identified a significant security risk that conflicted with product delivery timelines. How did you communicate the business impact and work with the team to reach a resolution?
Adapting to Rapidly Evolving AI Domains
What to coverThe AI security landscape changes quickly. Describe a situation where you had to learn a new AI framework or security threat vector on the fly to solve a critical problem. How did you validate your understanding?
02
Sharpen your knowledge
Topics to prepare
AI Red Teaming and Adversarial Testing
Review methodologies for identifying prompt injection, jailbreaks, and RAG poisoning. Exercise: Design a threat model for a RAG-based application and list specific adversarial test cases to validate guardrail effectiveness against indirect prompt attacks.
Secure AI Architecture and Mitigation Patterns
Focus on implementing secure prompt design, input/output validation, and grounding checks. Exercise: Draft a technical document outlining a mitigation strategy for preventing data leakage in an LLM-powered enterprise agent.
Skills to revise
AI Security Automation
Develop scripts to automate vulnerability discovery or test generation for AI models. Exercise: Write a Python script that uses an LLM API to scan source code for common security vulnerabilities or to generate synthetic test cases for an API endpoint.
Application Security Testing (SAST/DAST)
Integrate security validations into CI/CD pipelines. Exercise: Configure a sample pipeline to include automated SAST or DAST scans and define how to interpret findings specifically for AI-led microservices.
Threat Modeling for AI Systems
Apply threat modeling to AI-specific components like vector databases and model orchestration layers. Exercise: Create a data flow diagram for an AI agent and identify potential trust boundaries and attack vectors at each stage of the retrieval process.
04
Your preparation checklist
Develop an AI Security Red Teaming Case Study
Prepare a technical walkthrough of a project where you identified and mitigated risks in a GenAI or RAG-based application. Focus on how you tested for prompt injection, data leakage, or insecure tool usage, and explain the specific remediation patterns you implemented.
Demonstrate AI-Powered Security Automation
Create a brief summary or code snippet demonstrating how you have built or utilized AI-assisted tools for vulnerability discovery, automated code analysis, or security testing within a CI/CD pipeline.
Map Experience to AI Security Frameworks
Prepare to discuss your familiarity with OWASP Top 10 for LLMs and how you apply threat modeling to AI-specific architectures, such as vector databases, model interaction layers, and agentic workflows.
Prepare Evidence of Cloud-Native Security
Be ready to explain your approach to securing AI workloads in cloud environments (Azure, AWS, or GCP), specifically addressing container security (Kubernetes/Docker), identity management, and secrets handling in an AI-integrated SDLC.
Formulate Strategic Questions for the Interviewer
Prepare insightful questions regarding the team's current AI security roadmap, how they balance rapid AI adoption with risk governance, and the specific challenges they face in securing their internal RAG or agentic platforms.
Availability has not yet been verified. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy
Share this opportunityKnow someone who would be a great fit?