AI Architect – Agentic Systems (LLM & Multi-Agent Solutions)

About the role

We are seeking a hands‑on AI Architect to design and lead production‑grade multi‑agent systems that integrate large language models (LLMs) with enterprise data, APIs and business workflows. This role focuses on building robust, scalable solutions rather than simple chatbots or prompt engineering.

Key responsibilities

• Design multi‑agent architectures, including task decomposition, orchestration and coordination patterns.
• Define interaction models between LLM‑powered agents and enterprise data platforms, APIs and operational tools.
• Lead projects from proof‑of‑concept to production, addressing model selection, cost, security, privacy and responsible AI safeguards.
• Establish observability, tracing, feedback loops and control mechanisms for agent behavior.
• Design memory strategies (short‑term, long‑term, contextual grounding) and integrate agents with core systems.
• Collaborate with data, platform and engineering teams to embed AI capabilities across the organization.
• Mentor and train teams on best practices for scalable, reliable AI systems.

Required profile

• Strong background in software architecture and distributed systems.
• Hands‑on experience building complex LLM‑based applications and orchestration workflows.
• Deep understanding of Retrieval‑Augmented Generation (RAG) architectures and retrieval quality.
• Solid knowledge of LLM fundamentals, transformer architecture, training paradigms and limitation mitigation.
• Experience with multi‑agent frameworks (e.g., LangChain, Semantic Kernel, CrewAI) and generative AI lifecycle platforms (Amazon Bedrock, Google Vertex AI, Azure AI).
• Familiarity with relevant protocols (MCP, UCP, A2A, AP2) and concepts such as tool use, function calling, agent coordination, and memory management.

Required skills

• Software architecture
• Distributed systems
• LLM‑based application development
• RAG and retrieval optimization
• Transformer models, attention mechanisms, tokenization
• Prompt engineering, hallucination mitigation
• NLP concepts: semantic similarity, embeddings, vector stores
• Multi‑agent frameworks (LangChain, Semantic Kernel, CrewAI)
• Generative AI platforms (Amazon Bedrock, Google Vertex AI, Azure AI)
• Protocols: MCP, UCP, A2A, AP2
• Cloud platforms: AWS, Azure, GCP