Agentic AI Architect (Solutions Engineer)/Ins/10+ yrs exp/PERM

AI Architect / Lead AI Solutions Engineer

The AI Architect/ Lead AI Solutions Engineer will spearhead the technical direction, scaling, and governance of our production-grade Enterprise Agentic RAG platform and multi-agent ecosystems within the Financial/Insurance space.

This is an elite builder-leader mandate. You will bridge the gap between high-level business strategy and deep technical execution, moving beyond basic semantic pipelines to architect self-correcting, iterative reasoning loops. You will directly lead and mentor a dedicated squad of high-performing AI/ML engineers, maintaining absolute accountability for system resilience, cost optimization, and institutional compliance in a highly regulated environment.

Key Responsibilities

1. Production-Grade Agentic Architecture & Workflows

• Design and ship end-to-end multi-agent products utilizing state-driven, cyclic workflows via LangGraph, LangChain, or LlamaIndex.
• Act as the ultimate technical strategic leader for determining core infrastructure tools, orchestrators, and vector engines. Drive the macro-level selection, benchmarking, and deployment of optimal model strategies—strategically balancing commercial APIs (OpenAI, Anthropic) with fine-tuned, quantized, or localized open-source models (Llama, Mistral, DeepSeek) based on latency, cost, hardware constraints, and strict compliance/security protocols.
• Transition legacy data processes into dynamic, iterative reasoning loops (incorporating query decomposition, self-reflection, and real-time context validation).
• Identify, benchmark, and deploy optimal model strategies—balancing commercial APIs (OpenAI, Anthropic) with fine-tuned, localized open-source models (Llama, Mistral) based on latency and security protocols.

2. Enterprise-Scale Retrieval & Systems Engineering

• Architect high-precision, layout-aware semantic chunking pipelines tailored for complex insurance policies, financial tables, and legacy document structures.
• Implement production-grade hybrid search (combining dense vectors, sparse BM25 keyword matching, and Reciprocal Rank Fusion) integrated with two-stage cross-encoder reranking layers.
• Ensure structural scalability and high availability using advanced containerization (Docker, Kubernetes) and inference server optimizations (vLLM, PagedAttention).

3. Cost, Token & Performance Optimization

• Drive strict LLM unit economics at scale by implementing semantic caching, context-window compression, and tactical context budgeting.
• Architect dynamic, cost-based model routing layers to delegate low-complexity lookups to lightweight models while reserving frontier models for deep reasoning workflows.

4. AI Governance, Safety & Guardrails

• Deploy robust enterprise safety nets to eliminate hallucinations and secure tool execution environments.
• Enforce institutional compliance, data privacy protocols (automated PII masking/redacting), and Source Access Control Lists (ACLs) within data ingestion streams.
• Build automated LLM-as-a-judge evaluation frameworks (e.g., Ragas, TruLens) to meticulously track Faithfulness, context precision, and latency SLAs.

5. Technical Leadership & Organizational Design

• Directly manage, inspire, and set rigorous code/architecture standards for a team of specialized AI/ML and software engineers.
• Articulate complex, multi-agent concepts and technological trade-offs clearly to C-suite stakeholders, regulators, and non-technical business leaders.

Technical Stack & Requirements

• Orchestration & Agents: Expert-level, production-vetted mastery of LangGraph(critical), LangChain, or LlamaIndex for complex state-tracking and multi-agent coordination.
• Infrastructure & Vector DBs: Deep experience with enterprise vector databases (Pinecone, Milvus, Qdrant, pgvector) and enterprise-grade data platforms (e.g., Azure DevOps, AWS).
• Core Software Engineering: Mastery of Python, asynchronous programming, microservices frameworks (FastAPI), and LLMOps/observability toolsets (LangSmith, Weights & Biases).

Experience & Qualifications

• Total Experience: 10 to 15 years of robust software, data, or system architecture experience within complex, enterprise-scale environments (ideally Insurance, Banking, or highly regulated FSI).
• AI Leadership: A minimum of recent years operating as a hands-on technical direction leader or Principal AI Engineer, with a definitive track record of directing engineering squads to ship production-ready GenAI/Agentic systems (not just experimental PoCs).
• Education: Graduate degree in Computer Science, Software Engineering, or a heavily quantitative field (or equivalent deep industry experience).

Argyll Scott Asia is acting as an Employment Agency in relation to this vacancy.