About the Role :

We are looking for an AI Architect & Consultant to shape production-grade Generative and Agentic AI solutions for R&D initiatives. The role spans solution architecture, client consulting, and technical leadership, from understanding business needs and defining scalable, secure AI architectures to guiding decisions across models, agents, retrieval, integrations, and evaluation. You will work closely with stakeholders and engineering teams to balance business value, cost, risk, and time to market while providing clear technical direction.

Key Responsibilities :

- Own the consulting mandate end to end - current-state assessment, target-state ecosystem blueprint and multi-track roadmap - and the quality of every artefact presented to the client.

- Lead the architecture of agentic AI systems, including planning, reasoning loops (such as ReAct and plan-and-execute), tool use, memory, and reflection.

- Design multi-agent systems, including orchestration patterns, agent-to-agent handoffs, shared state, and task delegation, and study their failure modes.

- Define how agents connect to tools, data, and other agents using Model Context Protocol (MCP) and comparable standards, including secure tool permissions and sandboxed code execution.

- Address current-state gaps with AI-native capabilities and evidence each recommendation with a cost-benefit model and a staged investment case.

- Define the KPI and value-realisation framework (provisioning lead time, straight-through processing, effort reduction, observability coverage) and agree baselines with the client.

- Design human-in-the-loop checkpoints, so that high-risk actions require approval before an agent carries them out.

- Build and lead evaluation frameworks for agents, measuring task success, step-level and trajectory-level correctness, tool-call accuracy, cost, latency, and regression across model and prompt versions.

- Define guardrails and safety approaches for agents, including prompt injection defenses, output validation, action limits, and red-teaming.

- Set up observability for agent systems, including tracing of each reasoning step and tool call, so that failures can be diagnosed and reproduced.

- Lead the architecture of LLM-based components, including model selection, fine-tuning and adaptation (such as LoRA and instruction tuning), prompt and context design, and retrieval-augmented generation (RAG).

- Investigate and benchmark new models, agent frameworks, and research papers, and translate promising results into prototypes and production-ready designs.

- Set standards for testing AI-generated code and AI system behavior, including automated agent evaluation in CI/CD pipelines.

- Lead, mentor, and grow a team of AI engineers and researchers, setting clear expectations and providing regular feedback.

- Present research findings and architecture decisions to technical and executive audiences.

- Collaborate with product, delivery, and client teams to turn R&D findings into solutions.

Required Skills and Experience:

- Hands-on architecture experience with LLM applications in production or near-production: RAG, embeddings, vector databases and prompt orchestration frameworks.

- Agentic AI design - tool calling, multi-agent orchestration, memory and context management, planning, and systematic evaluation of agent behaviour.

- Data architecture depth: data modelling, pipelines, log and telemetry ingestion, semantic layers, metadata management and data governance.

- Proven experience designing and running agentic AI systems, including single-agent and multi-agent architectures.

- Hands-on experience with tool/function calling, MCP or similar tool-integration patterns, and agent orchestration frameworks (such as LangGraph, AutoGen, CrewAI, or equivalent).

- Experience with agent memory (short-term and long-term), state management, and context engineering.

- Cloud AI and data platform experience on GCP (Vertex AI, BigQuery) and/or the Azure / AWS equivalents.

- Experience building evaluation frameworks for agents and LLM outputs, including automated and human-in-the-loop methods.

- Experience with RAG, embeddings, and vector search.

- Strong Python skills and experience with ML frameworks (such as PyTorch or Hugging Face).

- Ability to read and critically assess current AI research, and to turn it into working prototypes.

- Prior experience leading and managing engineers or researchers.