AI Architect - GenAI & Agentic Systems

Drive Innovation in GenAI and Automation Role Overview We are looking for an AI Architect/Lead (Grade 140) to design, build, and deploy automation and Generative AI solutions end to end, from architecture and prototypes to production systems with strong post-deployment monitoring. You will work with cross-functional teams to drive agentification and shape platform capabilities used across the organization. Key Responsibilities Architect end-to-end AI and automation solutions on cloud-native platforms, from design through production. Build agentic systems: LLMs and other Generative AI models (diffusion, audio/video), RAG pipelines, and multi-agent orchestration with frameworks such as LangChain, LangGraph, and MCP. Build agent harnesses: tool orchestration, memory, verification loops, and guardrails such as permissions, sandboxing, and human approval for risky actions. Develop automation frameworks, and apply prompt and context engineering (retrieval, memory, context compaction) to improve accuracy, grounding, and relevance. Deploy through scalable CI/CD pipelines on Azure and/or AWS, with safe rollout and rollback. Monitor after release, with observability for traces, logs, latency, errors, token usage, and cost. Measure and improve through evaluation-driven development: quality and agent metrics, task suites, LLM-as-judge, regression tests on every prompt, model, or tool change, and benchmarking of cloud AI services. Partner with engineering, product, and business stakeholders on prototypes and demos, and drive adoption. What We’re Looking For Core Technical Skills Strong proficiency in Python and hands-on development experience. Hands-on experience with Azure and/or AWS. Experience building backend services and APIs, with automated tests around them. Experience with CI/CD and safe rollout of services, including secrets management, identity, and environment promotion. Generative AI & Advanced Capabilities Practical experience with Generative AI, including LLM-based systems. Strong understanding of RAG architectures, embeddings, vector databases, and grounding strategies. Experience with LangChain, LangGraph, or similar GenAI orchestration frameworks. Experience with AI agents, multi-agent systems, and protocols such as MCP and A2A. Experience with LLM and agent observability and evaluation (LangSmith, OpenTelemetry, or similar): quality metrics, agent metrics, traces, cost, and latency. Understanding of agent failure modes and defences: hallucination, prompt injection, and context rot. Nice-to-have skills Solid understanding of SQL and data handling for analytics or AI workflows. Hands-on experience deploying LLM and agent services on Azure OpenAI or a comparable platform, including structured outputs and versioned prompts or models. Hands-on experience with multi-modal AI (speech, transcription, synthesis, or related domains). Prompt or program optimisation (e.g. DSPy) and LLM fine-tuning (LoRA, SFT).