We’re looking for a Tech Lead specializing in Agentic Gateway Architecture to lead the design and implementation of a secure, scalable integration layer for AI agents, tools, and multi-agent workflows. The focus is on MCP tool integration, semantic tool discovery, A2A communication, identity federation, authorization, and gateway architecture within an enterprise AI platform.

The project involves building an enterprise-grade AI Agent Platform from scratch , providing a standardized foundation for developing, deploying, orchestrating, securing, and observing AI agents across multiple teams and use cases. The platform will cover the full agent lifecycle, including orchestration, observability, security, governance, integrations, and evaluation.

This is a greenfield initiative , giving engineers the opportunity to shape the architecture and make key technical decisions rather than maintain an existing solution. The client is a large integral enterprise investing heavily in modern cloud, data, and AI capabilities to support secure, scalable, and highly automated solutions.

Requirements:




- Model Context Protocol (MCP) — server aggregation, tool discovery, tool invocation lifecycle
- Agentic workflow patterns — agent-to-tool, agent-to-agent communication (A2A), multi-agent orchestration
- Semantic tool discovery — embedding-based search for agents to find tools by capability/intent
- Agentic authorization — policy enforcement on tool calls and agent interactions (Cedar policy language)
- Agent identity federation — JWT-based cross-agent authentication (Entra Agent ID or equivalent)
- AWS Bedrock AgentCore Gateway — MCP target registration, HTTP target routing, inference target configuration

Experience:
- 5+ years building AI/ML platforms or agentic systems
- Designed tool integration layers for AI agents (function calling, tool use)
- Implemented Model Context Protocol or similar agent-tool protocols
- Enterprise security reviews for AI systems (authorization, data acce

📌 Tech Lead (Sevilla)
🏢 Intellias
📍 Sevilla