Back to Join

Agentic AI / MCP & A2A Specialist

Join Our Team as an Agentic AI / MCP & A2A Specialist

Are you excited by production-grade agentic AI systems, protocol integration, model/provider portability, evaluation, telemetry, workflow orchestration, and operational reliability?

PortBlueSky is supporting a major multinational enterprise in scaling a widely adopted open-source platform for agentic AI and enterprise automation. We are looking for an Agentic AI / MCP & A2A Specialist to drive the platform’s AI capabilities across workflow orchestration, protocol integration, model abstraction, telemetry, evaluation, and production operations.

This role is for someone who understands that agentic AI in production is not just prompt engineering. It requires protocol design, authentication, tool safety, evaluation, deterministic checks, observability, retries, fallback strategies, governance, and workflows that can be audited and improved over time.

Role and Responsibilities

As Agentic AI / MCP & A2A Specialist, you will:

  • Drive production-grade agentic capabilities: Design and improve agentic workflows, protocol integrations, model/provider abstractions, telemetry, evaluation, and reliability patterns.

  • Own MCP and A2A integration patterns: Work with MCP transports, streamable HTTP/SSE, OAuth-based authorization, token handling, tool discovery, safe tool execution, A2A task/message semantics, agent cards, streaming behavior, polling, gateways, and external execution engines.

  • Improve model/provider portability: Support integrations across LLM-provider APIs, Bedrock, enterprise model gateways, vendor-agnostic abstractions, fallback logic, and multi-model switching.

  • Build evaluation and governance into workflows: Add deterministic checks, structured validation, task-specific metrics, regression tests, observability-driven evaluation, escalation paths, human-in-the-loop review, and checks-and-balances beyond simple “LLM as judge” patterns.

  • Create durable workflow solutions: Use orchestration platforms such as Argo Workflows, and potentially n8n, Temporal, Airflow, Dagster, or similar systems, to make workflows repeatable, auditable, and resilient.

  • Own production concerns: Improve telemetry, token usage accounting, provider fallback, memory/session handling, evaluation harnesses, continuous improvement loops, and operational visibility.

  • Build use-case and reference layers: Create sample integrations, marketplace components, Helm charts, and reference deployments for tools such as Langfuse, OpenTelemetry/OTLP, Phoenix, RAG backends, MCP servers, workflow engines, and evaluation/observability stacks.

  • Work across engineering boundaries: Collaborate with platform, system, and product engineers to avoid AI-layer isolation and ensure the platform remains portable across providers, clouds, identity systems, telemetry stacks, workflow engines, and client infrastructure constraints.

What We Offer

  • A cutting-edge AI platform challenge: Work on production agentic AI infrastructure, not throwaway prototypes.

  • A senior technical culture: PortBlueSky publicly emphasizes challenging enterprise projects, top-tier experts, high communication standards, and open-source involvement.

  • Remote-first working conditions: Work flexibly in a distributed team with a strong focus on autonomy and quality.

  • Real production impact: Help shape how agentic AI systems are evaluated, governed, monitored, integrated, and deployed in large enterprise environments.

  • Room to specialize and connect disciplines: Combine AI, protocols, workflow orchestration, Kubernetes, platform engineering, and systems design.

About You

You are an experienced AI/platform engineer with:

  • Production-level hands-on experience with MCP, A2A, OAuth-based MCP authorization, token handling, agentic routing, LLM-provider APIs, Bedrock, and vendor-agnostic model abstraction.

  • Deep knowledge of MCP and A2A integration patterns, including transports, streaming, tool discovery, task semantics, external execution engines, and gateway patterns.

  • Strong understanding of production AI workflow design, including verification, evaluation, governance, observability, fallback, and escalation.

  • Experience building durable, auditable, repeatable workflow solutions.

  • Ability to work across Go and Python, with enough TypeScript familiarity to understand broker, CLI, and dashboard implications.

  • Strong communication skills and the ability to work closely with platform and system engineers.

  • Fluent English and comfort working in an international, remote-first team.

  • EU residency and permission to work in the EU.

Nice to Have

  • Experience with Argo Workflows, n8n, Temporal, Airflow, Dagster, or similar orchestration systems.

  • Experience with Langfuse, Phoenix, OpenTelemetry/OTLP, RAG backends, MCP servers, evaluation frameworks, or observability stacks.

  • Strong enough Kubernetes/platform engineering experience to support deployment and productionization work.

  • Strong enough Go/controller-runtime experience to contribute to core runtime or Kubernetes-native system components.

  • Experience with enterprise model gateways, cloud AI platforms, managed identity, private networking, or regulated production environments.

  • Open-source contribution experience or developer-facing documentation experience.

Technologies & Tools You May Work With

  • MCP

  • A2A

  • OAuth

  • LLM-provider APIs

  • Bedrock

  • Vendor-agnostic model gateways

  • Go

  • Python

  • TypeScript

  • Argo Workflows

  • n8n / Temporal / Airflow / Dagster

  • Langfuse

  • Phoenix

  • OpenTelemetry / OTLP

  • RAG backends

  • Helm

  • Kubernetes

  • Enterprise identity and telemetry systems

Ready to build agentic AI systems that survive real enterprise use? Apply now and help shape production-grade AI automation at scale.

Apply Now!