New opportunity

Head of AI Engineering & Cyber Solutions

Societe Generale · Bangalore, India

About this role

Lead AI engineering and cyber solutions for Societe Generale’s Bangalore, India, AI Hub. This hybrid, permanent role combines hands-on technology leadership with responsibility for delivering secure, production-grade AI solutions for cybersecurity teams and building a global centre of excellence.

Responsibilities

Define AI strategy across security, risk and cyber enablement. Lead engineers, developers, data engineers and automation specialists in designing, building, deploying and supporting AI-enabled platforms, tools and applications. Over time, take on broader IT development delivery beyond AI use cases.
Translate cyber requirements into scalable solutions for security operations, threat detection, vulnerability management, application security, automation and reporting. Guide architecture, platform choices, testing, releases, DevSecOps, observability and production readiness.
Establish security practices for AI/ML pipelines, GenAI and LLM systems, agentic applications and cloud AI components. Assess AI use cases throughout their lifecycle; define guardrails, controls and standards aligned with enterprise governance, regulatory and audit expectations.
Build the team’s AI engineering and security capabilities, coordinate with global security, data and technology leaders, and establish processes for use-case assessment, prioritization, validation and risk reporting.

Required background:

At least 15 years of progressive experience in software engineering, IT development, platform engineering or technology delivery, including significant leadership responsibility. Proven delivery of enterprise-grade applications, platforms, automation, data or AI solutions, with hands-on AI/ML/GenAI experience and a solid understanding of cybersecurity. A career-long cybersecurity specialization is not required.
Expertise in enterprise application architecture, APIs, cloud-native engineering and secure software delivery. Understand LLM applications, RAG, agentic workflows, vector databases, prompt engineering, model evaluation, MLOps, AI guardrails and AI threats such as prompt injection, jailbreaking and data-security risks. Bring knowledge of cloud AI and security across Azure, AWS or GCP, and data-platform security, including Databricks or Snowflake. Be able to influence senior stakeholders in a global, matrixed environment.

Preferred

Experience in financial services, banking, cybersecurity, risk or another regulated industry; Python or Java, with LangChain or LangGraph advantageous; and CISSP, CISM, CISA, GIAC or AI/ML security certifications or equivalent expertise.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

AI engineeringCybersecuritySoftware engineering leadershipGenerative AIMachine learningLarge language modelsRAGAgentic AIAI securitySecurity operationsThreat detectionVulnerability managementApplication securityCloud securityCyber riskDevSecOpsSoftware architectureAPIsMicroservicesCloud-native engineeringEnterprise integrationAgileCI/CDAutomated testingObservabilitySecure codingPythonJavaData engineeringData pipelines

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Build and lead the AI Hub
The role is tasked with defining AI strategy across security, risk and cyber enablement, scaling the India hub as a global centre of excellence, and leading AI, software, data and automation engineers.
Deliver practical AI for cyber teams
The posting calls for translating cyber needs into AI-enabled tools and platforms for security operations, threat detection, vulnerability management, application security, automation and reporting.
Own production-grade engineering
Accountability spans problem framing through deployment, adoption and support, with emphasis on architecture, testing, releases, DevSecOps, observability and maintainable integration with enterprise systems.
Secure AI throughout its lifecycle
The role establishes controls for AI/ML pipelines, LLM and agentic applications, cloud AI components and data platforms; assesses use-case risk; and aligns solutions with governance, regulatory and audit expectations.
Establish a global operating model
The posting assigns processes for use-case assessment, prioritization, validation and risk reporting, alongside collaboration with global security, data and technology leaders. It also anticipates broader IT development ownership over time.

Likely challenges to think through

Choosing useful cyber use cases
Likely preparation challenge, not a stated employer problem: attractive AI ideas may differ in cyber value, data readiness and risk. Prepare a way to compare use cases, validate outcomes with cyber teams and decide which merit production investment.
Moving GenAI into dependable operations
Likely preparation challenge: rapid experimentation can conflict with reliability, maintainability and production support. Prepare a concrete delivery approach covering evaluation, automated tests, observability, releases and ownership after deployment.
Balancing AI capability with controls
Likely preparation challenge: LLM and agentic functionality can introduce prompt-injection, data-security and hallucination risks while teams seek timely delivery. Prepare to explain how lifecycle assessments and embedded guardrails would shape design decisions.
Scaling while aligning global stakeholders
Likely preparation challenge: building an India-based capability while coordinating across a global matrix may create competing priorities, particularly as the remit expands beyond AI. Prepare an approach to decision rights, prioritization and risk reporting without assuming a specific existing process.

Practice your answers

Technical practice

How would you turn a cybersecurity need into a production-grade AI use case?

What to coverSuggested preparation: Explain how you would frame the problem with cyber stakeholders, establish a baseline and success measures, and decide whether AI adds value over simpler automation. Cover validation, security review, adoption and ongoing support.

How would you design an enterprise AI application for cybersecurity teams?

What to coverSuggested preparation: Discuss APIs, integration, cloud-native components and choices that support scalability and maintainability. Explain how you would assess platform trade-offs while keeping the solution secure and observable.

When would you use RAG for a cyber use case, and how would you evaluate it?

What to coverSuggested preparation: Address source quality, access controls, retrieval, vector databases and prompt design. Compare answer quality and grounding with latency, cost and the risk of exposing sensitive information.

What controls would you put around an agentic AI workflow?

What to coverSuggested preparation: Discuss permissions, tool access, approval boundaries and monitoring for an agent that can act on cyber information. Explain how you would test for prompt injection, unintended actions and failures across orchestration steps.

How would you secure an AI or ML pipeline throughout its lifecycle?

What to coverSuggested preparation: Cover data provenance and protection, model and dependency controls, evaluation before release, and monitoring after deployment. Explain where guardrails belong in the solution and how lifecycle risks would be reassessed.

How would you evaluate an AI-assisted threat-detection or security-operations tool?

What to coverSuggested preparation: Compare it with the existing workflow using measures such as detection quality, false positives and analyst effort. Address how analysts would verify outputs and what evidence would justify scaling or stopping the use case.

What security trade-offs would you assess when choosing cloud AI and data-platform components?

What to coverSuggested preparation: Discuss identity, data access, isolation, logging and integration for AI workloads on Azure, AWS or GCP and data platforms such as Databricks or Snowflake. Explain how enterprise standards and the use case would shape the choice rather than assuming one platform is required.

How would you make an AI-enabled application ready for release and production support?

What to coverSuggested preparation: Cover automated testing, CI/CD, DevSecOps, observability and a support plan for both application and model behaviour. Explain how you would handle a degradation or unsafe output after release.

How would you prioritise and govern a portfolio of AI-for-cyber use cases?

What to coverSuggested preparation: Propose assessment criteria for cyber value, feasibility, validation effort and risk. Describe how controls, decision records and risk reporting could support enterprise governance and audit expectations.

How would you set an AI engineering strategy while preparing to own broader IT delivery?

What to coverSuggested preparation: Explain how you would sequence reusable engineering capabilities and cyber use cases without letting strategy outrun delivery. Address how architecture and quality standards could extend to non-AI development as the remit grows.

Behavioural practice

Tell me about a time you built or scaled a multidisciplinary engineering capability.

What to coverSuggested preparation: Use a real situation, your actions and the result. Focus on how you defined capabilities, responsibilities and ways of working across different technical disciplines.

Describe a time you aligned senior stakeholders with conflicting technology or risk priorities.

What to coverSuggested preparation: Use a real situation, your actions and the result. Explain how you surfaced trade-offs, influenced a decision and maintained collaboration across teams or regions.

Tell me about a time you challenged an architecture or platform choice.

What to coverSuggested preparation: Use a real situation, your actions and the result. Show how you tested the choice against delivery, security and maintainability needs and handled disagreement.

Describe a time a production solution did not perform as intended.

What to coverSuggested preparation: Use a real situation, your actions and the result. Explain how you coordinated diagnosis and support, communicated risk, and changed engineering practices afterward.

Tell me about a time you helped a team adopt a new technical capability.

What to coverSuggested preparation: Use a real situation, your actions and the result. Focus on how you developed skills, worked with partner teams and assessed whether the capability improved delivery.

Sharpen your knowledge

Topics to prepare

Enterprise AI engineering architecture
The posting emphasizes APIs, microservices, cloud-native integration and production-grade delivery. Revise by explaining the trade-offs between two architectures for an AI-enabled cybersecurity application, including scalability, maintainability and production support.
LLM, RAG and agentic patterns
The role calls for hands-on understanding of RAG, vector databases, orchestration, agentic workflows, model evaluation and MLOps. Revise by tracing a use case from data ingestion through retrieval or tool use, evaluation and ongoing model lifecycle management.
Cybersecurity use cases
The posting names security operations, threat detection, vulnerability management, application security, automation and reporting. Revise by comparing two of these domains: identify the cyber team's problem, the data an AI solution would need and how its output would be validated.
AI threats and secure adoption
The role explicitly covers prompt injection, jailbreaking, hallucinations, data security, adversarial robustness and agentic AI security. Revise by mapping each threat to a plausible failure mode and a preventive or detective control.
Cloud and data-platform security
The posting calls for AI workload security across Azure, AWS or GCP and security considerations for Databricks or Snowflake. Revise by drawing the trust boundaries of one cloud AI workload and identifying where data access, logging and guardrails need review.
Governance and engineering leadership
The role involves building the India AI Hub, coordinating global teams and defining use-case assessment, prioritization and risk reporting. Revise by outlining decision criteria for three competing cyber AI proposals and how you would explain the trade-offs to senior stakeholders.

Skills to revise

Design a cyber AI solution
Practise producing a one-page design for an AI-assisted vulnerability-management workflow: show APIs, data flow, integration points, failure handling and production support. This exercises the end-to-end architecture and cyber-requirements translation described in the posting.
Build and evaluate a small RAG workflow
As a preparation exercise, implement a small retrieval-backed question-answering workflow using non-sensitive sample cyber documents; test retrieval quality, answer grounding and failure cases. The posting explicitly seeks hands-on RAG, prompt engineering and model-evaluation understanding; this exercise is not a specified employer project.
Threat-model an AI workflow
Create an attack-and-control table for a hypothetical agent that reads security findings and calls approved tools. Test prompt-injection and unauthorized-data-access scenarios, then specify guardrails and evidence of their effectiveness. This reflects the posting's AI security and agentic-risk responsibilities.
Prepare a production-readiness review
Draft a release checklist for an AI-enabled service covering automated tests, CI/CD, DevSecOps checks, observability, rollback and run support. The role explicitly owns engineering practices and production readiness.
Present a use-case decision
Write a short decision brief that ranks three hypothetical security-operations AI use cases by expected value, validation effort and risk; include a recommendation and open questions for security and technology leaders. This practises the posting's prioritization, risk visibility and global stakeholder responsibilities.

Your preparation checklist

Map evidence to the role
Prepare a one-page evidence map for the stated 15+ years of technology delivery, significant leadership, hands-on AI delivery, and cybersecurity understanding. Mark any regulated-industry experience or certifications as preferred, not required.
Rehearse an end-to-end delivery walkthrough
Select a real enterprise AI or software platform project, if available. Explain your decisions from problem framing through architecture, testing, deployment, adoption, observability, and production support; distinguish your own contributions from the team’s.
Sketch an AI-for-cyber design
As a practice exercise, diagram a production-ready solution for one listed use case, such as vulnerability management or security-operations reporting. Show data flows, APIs, cloud and platform choices, evaluation, integration, and operational trade-offs.
Prepare an AI security review
For that hypothetical design, write a short threat-and-control table covering data access, prompt injection, jailbreaks, hallucinations, agent permissions, and model or pipeline risks. Explain where guardrails, testing, monitoring, and human review would fit.
Outline a use-case governance process
Draft a proposed way to assess, prioritize, validate, and report risk for AI use cases across their lifecycle. Identify what evidence would support enterprise governance and audit review; present this as your proposal, not Societe Generale’s existing process.
Develop a team-building narrative
Prepare examples, where available, of leading engineers and technical leads, improving delivery practices, and influencing global stakeholders. Add a proposed approach to building the India AI Hub’s engineering and security capabilities.
Write targeted employer questions
Ask which cyber use cases are highest priority, how global security and technology teams share decisions, what AI controls and platforms already exist, and how success and the possible expansion into broader IT delivery would be measured.

Availability last checked: 4 Nov 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in India. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…