Technical Architect

Job Specification: Lead AI Architect (Technical & Implementation Focus)

Sector: Insurance

Location: London (Hybrid: 1–3 days in-office/client site)

Duration: 12 Months Initial (Long-Term Expectation)

IR35 Status: Expected Outside IR35

Start Date: Early January 2026

Role Overview

We are seeking a heavyweight Technical AI Architect to lead a major Gen AI transformation for a global Insurer. This is a \"player-coach\" style architectural role.

While you will define the strategic direction for an organization at low AI maturity, you must possess the technical depth to design specific implementation patterns and provide granular guidance to engineering teams. You will not just \"recommend\" RAG; you will define the orchestration logic, the embedding strategies, and the technical guardrails required to make it production-ready.

Critical Requirement: You must be able to articulate why specific technologies (e.g., specific LLMs, vector databases, or orchestration frameworks) were chosen in your past projects and explain the flow of data at a granular level.

Key Responsibilities

1. Technical Solution Design & Deep-Dive

• Architectural Blueprints: Move beyond high-level boxes and arrows to create detailed designs for RAG patterns, Agentic workflows, and Graph-based retrieval systems.
• Orchestration & Pipelines: Define the execution logic for AI pipelines, including prompt chaining, state management in agentic systems, and integration with middleware/APIs.
• Data & AI Ecosystem: Ensure seamless integration between Gen AI components and the existing enterprise stack (Azure & Snowflake).

2. Engineering Leadership

• Implementation Guidance: Provide \"at-the-keyboard\" level guidance to engineering teams on prompting strategies, configuration, and model fine-tuning.
• Tooling Evaluation: Conduct rigorous, technical \"Pros vs. Cons\" evaluations of model families and frameworks (e.g., LangChain, LlamaIndex, Semantic Kernel) based on latency, cost, and accuracy.

3. Technical Governance & Security

• LLM Security: Implement deep-tier security patterns, including prompt injection mitigation, PII masking, and robust governance frameworks for LLM outputs.
• AI Ops (LLMOps): Establish repeatable patterns for model monitoring, versioning, and evaluation (evals) to move use cases from PoC to scalable production.

Required Experience & Technical Skills

Core Technical Depth (Non-Negotiable)

• Proven Implementation: You must be able to walk through the \"how\" of your previous projects—explaining the specific orchestration, memory management, and data flow used.
• Advanced Patterns: Deep experience in RAG, Agent-to-Agent communication protocols, and In-memory Graph RAG.
• The \"Glue\": Expert-level knowledge of enterprise integration, middleware, and how to embed AI into complex, legacy operational workflows.
• Cloud Platforms: Deep Azure expertise is required, but you must demonstrate \"platform-agnostic\" thinking with experience in at least one other major cloud provider.

Consultative Ability

• Maturity Uplift: Experience taking a \"low-maturity\" client and teaching them the difference between \"hype\" and \"deployable value.\"
• Clarity of Communication: Ability to present complex technical decisions without relying on slides. You should be able to white-board a solution and defend your architectural choices under technical scrutiny.

Interview Expectations

Candidates will be evaluated specifically on their ability to:

• Explain the \"Why\": Why this model? Why this specific vector search strategy?
• Detail the \"How\": Explain the technical execution of a pipeline or orchestration layer.
• Security Depth: Go beyond \"generic\" security; discuss specific implementation of guardrails and data privacy in an LLM context.