Be the one building AI-powered experiences where they matter most.
At Genesys, we help organizations create better customer experiences through AI-powered experience orchestration. Our platform connects people, systems, data and AI to help organizations deliver more personalized service, improve operational efficiency and build stronger customer relationships.
Help build, support and operate technology used by more than 8,000 organizations in over 100 countries – moving AI from possibility to production in real-world enterprise environments every day.
Principal Engineer, MCP & A2A
Department & Team: Engineering – MCP Orchestration
Location: Virtual Office (Portugal)
Summary
We're looking for a Principal Engineer to build the MCP, A2A, and agentic workflow systems at Genesys. This is a hands-on engineering role at the deepest technical level of the domain. You will design and ship the MCP server and client infrastructure, the agent-to-agent communication layer, and the orchestration runtime that agentic workflows run on. You are the person who solves the hardest problems in this stack, and the one other engineers come to when an agent system does something nobody can explain.
Key Responsibilities
Design and build Genesys MCP infrastructure: servers, clients, tool and resource schemas, transports, auth and scoping, versioning, and the internal MCP registry
Build the A2A layer: agent discovery, capability advertisement, task delegation and handoff, message contracts, trust boundaries, and failure semantics between agents
Build the agentic workflow runtime: planning and tool-call loops, state and context management, memory, human-in-the-loop checkpoints, retries, and termination conditions
Take the hardest problems in the stack end to end, from prototype through production
Build the evaluation and observability tooling for agent systems: tracing across agent hops and tool calls, eval harnesses, regression gates, cost and latency instrumentation
Engineer the security of agent systems: tool sandboxing, least-privilege tool access, prompt-injection and tool-abuse defenses
Debug and resolve the production failures nobody else can, across non-deterministic multi-agent paths
Make build-vs-buy calls on agent infrastructure: MCP gateways, agent runtimes, orchestration frameworks, eval tooling
Work directly with the engineers building on this stack, through code review, design review, and pairing
Minimum Requirements
12+ years of professional software engineering experience, still writing code
Deep hands-on expertise with MCP: has designed and shipped MCP servers and clients, not just consumed them
Deep hands-on expertise with A2A or equivalent multi-agent communication patterns, including delegation, handoff, and failure handling across agent boundaries
Proficiency in Java and/or TypeScript/Node
Has built and operated multi-step agentic workflows in production under real reliability, latency, and cost constraints
Experience building evaluation and observability for non-deterministic systems
Track record shipping platform or infrastructure other engineers build on
Excellent written and verbal communication, including explaining agent architecture tradeoffs to non-technical stakeholders
Desirable Skills
Contributions to the MCP or A2A specs, or to open-source agent orchestration projects
Prior Principal or Distinguished Engineer role building agent or platform infrastructure
Experience with agent security and adversarial testing of tool-using systems
Conference talks or published technical work on MCP, A2A, or agent orchestration
Working at Genesys
AI at enterprise scale – Build, support and operate AI-powered technology used by more than 8,000 organizations worldwide. 150+ new AI features were released in the last fiscal year.
A flexible-first culture – Join a global team of nearly 7,000 employees with flexible ways of working designed to help people do their best work.
Growth in the AI era – Build future-ready skills through mentorship,…