AI Automation Specialist_2881

Objectives of Job

The AI Generalist is a pivotal technical role within the 'Automation chapter' of the Security Services Tribe. The role is responsible for designing, developing, and operationalising Agentic AI solutions — encompassing LLM-based workflows, Retrieval-Augmented Generation (RAG) pipelines, and multi-step orchestration frameworks — that enable to apply intelligence on large volumes of security data at scale.

The engineer will work closely with security service product owners across Vulnerability Management, Web Application Firewall (WAF), Governance Risk and Compliance (GRC), and other domains to identify automation opportunities, build secure AI-powered workflows, and enforce robust AI guardrails appropriate for the sensitive cybersecurity environment.

Experience, Technical and Functional Skills

Core Responsibilities:

• Partner with security product teams (Vulnerability Management, WAF, GRC) to analyse operations, identify AI/automation opportunities, and translate them into production-grade agentic workflows.
• Design and develop Agentic AI systems using LLM frameworks (LangChain, LangGraph, AutoGen, or equivalent) to automate multi-step security workflows with minimal human intervention.
• Build and maintain Retrieval-Augmented Generation (RAG) pipelines to enable intelligent querying and contextual analysis of large security datasets, logs, and knowledge bases.
• Implement AI orchestration and workflow automation solutions that integrate with security tools such as SIEM, SOAR, vulnerability scanners, WAF platforms, and GRC systems via REST APIs.
• Apply security-first principles to all AI development — implement guardrails, prompt injection defences, output validation, and rate-limiting controls to protect AI pipelines in a sensitive cybersecurity context.
• Develop data analytics capabilities to derive actionable intelligence from large-scale security telemetry, alert data, and threat feeds using Python, SQL, and data visualisation tools.
• Ensure responsible AI practices: monitor for model drift, hallucination risks, bias in AI-driven decisions, and maintain audit trails for AI-assisted security actions.
• Collaborate with DevSecOps and platform teams to deploy AI models and workflows through CI/CD pipelines with appropriate access controls and secrets management.
• Continuously evaluate emerging LLM frameworks, agentic AI patterns, and AI security tooling to improve the ACDC's AI maturity and operational resilience.

Primary Skills:

• Agentic AI Development (Mandatory) — hands-on experience building multi-agent or agentic AI systems using LangChain, LangGraph, AutoGen, CrewAI, or comparable orchestration frameworks.
• LLM Integration (Mandatory) — practical experience integrating large language models (OpenAI GPT, Anthropic Claude, Mistral, or open-source equivalents) into production applications via APIs.
• RAG Pipelines (Mandatory) — design and implementation of Retrieval-Augmented Generation systems including vector databases (Pinecone, Weaviate, FAISS, ChromaDB), embedding models, and chunking strategies.
• Python (Mandatory) — advanced scripting for AI workflow automation, data processing, API integration, and security tooling.
• REST API Integration (Mandatory) — experience integrating AI agents with third-party security platforms (SIEM, SOAR, vulnerability management, ticketing) via REST APIs.
• AI Security & Guardrails — knowledge of agentic AI security risks (prompt injection, data leakage, tool misuse) and hands-on experience implementing guardrails, input sanitisation, and safe output handling.
• Data Analytics — proficiency in analysing and deriving insights from large structured/unstructured datasets using Python (Pandas, NumPy), SQL, and visualisation tools (Grafana, Power BI, or similar).
• Cybersecurity Domain Awareness — working knowledge of security operations concepts: vulnerability management lifecycles, WAF rule logic, GRC frameworks, SOC workflows, and security alert triage.

Good to Have:

• Experience with AI observability and monitoring tools (LangSmith, Weights & Biases, MLflow) for tracking LLM performance and detecting anomalies in agentic pipelines.
• Familiarity with vector search and knowledge graph technologies for enriching RAG systems with structured security intelligence.
• Cloud platform knowledge (AWS, Azure, or GCP) — particularly AI/ML services (Bedrock, Azure OpenAI, Vertex AI) and security-hardened deployment patterns.
• CI/CD tooling (Jenkins, GitHub Actions, GitLab CI) for automated testing and deployment of AI models and agentic workflows.
• Awareness of OWASP Top 10 for LLM Applications and MITRE ATLAS framework for AI threat modelling.