A hands-on principal-level architect who acts as chief Design Authority for agentic AI systems, designing agent frameworks, MCP context protocols, tool servers, and production guardrails. Works within a build team creating autonomous AI platforms for an international FinTech & Telecom client.
For our client – an international FinTech & Telecom group – we are building a world-class AI capability (AI Squad) from the ground up. We create autonomous and semi-autonomous agentic systems and intelligent platforms that will transform and automate financial services for millions of users.
We are seeking an Agentic & MCP Architect to serve as the chief Design Authority for agentic systems across Group FinTech. This is a hands-on architecture role within a build team – ranging from designing agent frameworks and tool servers, through defining context protocols (MCP) and integration patterns, to ensuring predictable, safe execution in production environments.
Requirements (Must-have):
Experience: Minimum 8–10 years of experience in software or machine learning engineering, including at least 3 years in a Technical Design Authority or Lead Engineering role.
Production Agentic Practice: Proven hands-on track record of designing and deploying agentic or workflow automation systems running at scale in production.
Domain Expertise: Deep practical knowledge of large language models (LLMs), agent frameworks, tool/function calling, context protocols (MCP), planning, and memory design.
Enterprise Integration: Experience integrating automated AI systems with core transactional platforms and enterprise applications.
Guardrails & Security: Experience designing permissioning, guardrails, and human oversight for systems taking action in regulated environments.
Nice-to-have:
Experience in FinTech, banking, or other high-volume transactional environments.
Recognized cloud, security, or AI engineering certifications.
Postgraduate qualification (Master's or Ph.D.) in a technical discipline (Computer Science, Data Science, Mathematics, or a related field).