Agentic System Architect

You will design system patterns, MCP interfaces, and cognitive context layers that allow LLMs and autonomous agents to safely refactor, test, and ship production code with minimal human intervention.

The candidate will have responsibilities across the following functions:

Agent-operable System Architecture:

• Decompose legacy monoliths into agent-readable, modular systems with strict boundaries and single responsibility.
• Lead incremental modernisation using the Strangler Pattern, wrapping legacy logic in modern, contract-driven interfaces.
• Enforce SOLID principles, Dependency Injection, and Hexagonal Architecture to ensure deterministic AI execution and low regression risk.

Agentic Framework and MCP Leadership:

• Architect and maintain the agent context layer: standardised Skills, Rules, and Commands for AI-driven engineering workflows.
• Build and operate Model Context Protocol (MCP) servers exposing legacy APIs, services, and databases as typed, secure, AI-consumable tools.
• Own contract-first API design as the primary interface between human intent and autonomous agent execution.

Engineering Governance and AI Quality Control:

• Act as architectural gatekeeper for AI-generated pull requests, ensuring scalability, security, and long-term maintainability.
• Mandate test-driven development (TDD) and characterisation testing to preserve legacy behaviour during refactoring.
• Monitor and optimise agentic reasoning loops to balance cost, speed, and architectural integrity.

Requirements:

• 15+ years in software architecture, platform engineering, or technical leadership.
• Proven experience modernising large, undocumented legacy systems.
• Deep hands-on expertise with TypeScript, . NET, and Node.js .
• Strong background in API design, distributed systems, and modular architectures.
• Practical experience with agentic development, MCP, LLM tooling, or AI-assisted engineering.