Senior Manager - Artificial Intelligence Information Security

Job Description

Requirements

• Minimum Qualifications
• Bachelor's degree in Computer Science/Engineering/Mathematics
• Honours degree advantageous
• Relevant security certifications (e.g., CISSP, CCSP), plus data/AI credentials (e.g., cloud AI specialties, ML engineering) are advantageous.
• Experience
• 5–10+ years across cybersecurity/platform security with 3-5+ years securing AI/ML platforms, GenAI/LLM ecosystems, or data-intensive analytics at enterprise scale.
• Demonstrable track record establishing security operating models/COEs and driving group-wide policy adoption in complex, multi‑country organizations (preferably telco/fintech).
• Hands‑on exposure to cloud-native AI stacks (in particular Azure), MLOps toolchains, and embedding controls in agile product delivery.
• Experience integrating AI systems into SOC/SIEM, designing AI incident response, and conducting AI red teaming and robustness testing.
• Core Competencies and skills
• Security architecture for AI/ML
• Threat modelling and adversarial ML
• Secure MLOps; data security & privacy
• GenAI/LLM guardrails
• Risk & compliance
• Leadership and stakeholder management

Responsibilities

• The Senior Manager Information Security AI COE is responsible for the following key performance areas:
• AI COE (Security Pillar) & Governance
• Establish the AI COE security mandate, operating model, and RACI across Group functions and OPCOs; embed BRAIN policy requirements into standards, procedures, and product gates.
• Support activities of AI Steerco, ISF, AWC and TSGC; integrate with Group Risk, Legal, Privacy, and Procurement for policy, third-party, and contract controls.
• Define enterprise AI control objectives, assurance plans, and attestation mechanisms aligned to BRAIN and Group policies (in particular GISP).
• Enterprise AI Security Strategy & Architecture
• Define and implement MTN's AI security strategy aligned to Group cyber reference architecture and recognized industry principles (e.g., Zero Trust, cloud security frameworks, secure SDLC, IAM, and detection & response).
• Publish secure architecture standards for AI platforms, MLOps stacks, and model-serving patterns; embed security-by-design across the AI lifecycle (data, training, evaluation, deployment, operations).
• Generative AI & LLM Security (BRAIN Policy Enablement)
• Operationalize BRAIN policy via guardrails: prompt security, input/output filtering, data loss prevention for prompts/outputs, access control & usage monitoring, and secure LLM architecture patterns for hosted and API models.
• Define approval workflows for model access, use cases, and sensitive data handling; implement usage analytics and model egress controls to prevent leakage.
• Adversarial Machine Learning Defence & AI Red Teaming
• Build MTN's AI threat modelling methodology, in collaboration with GIS; institute adversarial robustness testing (poisoning, evasion, prompt injection, API exploitation) and AI red teaming exercises for models and applications, in collaboration with GIS Cyber-defense.
• Define secure model deployment controls and post-deployment behavioural drift monitoring for manipulation detection in collaboration with the CCOE.
• Secure MLOps & Data Security
• Set Secure MLOps standards: CI/CD for models, integrated security testing, hardened model registry & artifact stores, signed/attested models, and pipeline authN/authZ, in collaboration with S2 COE and main relevant suppliers.
• Protect training/validation datasets; enforce patterns for secure data ingestion, sensitive data minimization, and prevention of training data leakage.
• AI Security Monitoring & Incident Response
• Integrate AI platforms, data pipelines, and model-serving endpoints with enterprise SIEM/SOC for continuous monitoring and anomaly detection.
• Extend cyber incident response playbooks to AI scenarios, in collaboration with Cyber-defense: containment of compromised models, forensic acquisition of model artifacts, and post-incident model integrity verification.
• Risk, Compliance, and Policy Alignment
• Map AI security controls to BRAIN and Group policy; align with POPIA, GDPR, ISO/IEC 27001/27701, ISO/IEC 42001 (AI), NIST AI RMF, and applicable sectoral obligations in telco and fintech.
• Establish Model Risk Management (MRM) with 1st/2nd line—including risk taxonomy, criticality tiers, control baselines, testing cadence, and assurance reporting to Group risk committees.
• Oversee third-party & cloud AI risk, in collaboration with Legal, Procurement and GIS-GRC: vendor due diligence, contract clauses (data, IP, security SLAs), and ongoing assurance.
• Platform & Product Enablement
• Partner with Connectivity, Fintech (MoMo), and Infraco product lines—and with the GenAI adoption stream—to embed design-time controls, privacy-by-design, and production guardrails into AI-enabled services.
• Define reference architectures and secure patterns for common use cases (RAG, copilots, fraud analytics, network optimization), in collaboration with GIS.
• People Leadership & Capability Uplift
• Build and lead a high-performing AI Security team within the AI COE; develop community of practice with GIS, OPCO CISOs and DPOs.
• Launch training & awareness for engineers, data scientists, and product teams on BRAIN policy, secure GenAI usage, and adversarial ML, in collaboration with GIS OPCO Operations.
• Performance, KPIs & Reporting
• Define and track KPIs: % AI use cases cleared by BRAIN gates, time-to-approve models, adversarial test coverage, model drift MTTR, policy exceptions, third-party assurance completion, and reduction in AI security incidents.
• Provide executive dashboards and Board/Exco reporting on AI risk posture and maturity.
• Budget, Tooling & Vendor Management
• Own budget for AI security tooling (e.g., secrets scanning for ML, model provenance/attestation, content safety, AI Security Posture Management), and manage vendors/partners via Group procurement.