Job Summary

• Design, implement and test Google CXAS conversational agents that understand customer requests, invoke the Universal Gateway and deliver accurate, natural responses through the selected text and voice channels.

• Configure and develop Google CXAS root and specialist agents for approved business journeys.
• Translate requirements into conversation flows, agent responsibilities, instructions and tool contracts.
• Implement the agreed session-start and initialization sequence, including when the Gateway must be called relative to the greeting.
• Configure tool calls so agents pass the correct session identifiers, journey information, user-provided values and requested operations to the Gateway.
• Consume Gateway responses containing state, permitted next actions, response guidance and business data.
• Keep authoritative business state and policy enforcement in the Gateway; avoid maintaining conflicting state definitions in agent instructions.
• Implement agent routing and handoffs while preserving the required conversation context.
• Configure clarification, fallback, escalation and unsupported-request handling.
• Implement voice behaviour, including microphone interaction, spoken responses and interruptions, for the agreed channel.
• Configure appropriate guardrails without blocking valid, authorized customer enquiries.
• Validate behaviour through native Canvas Preview and automated session tests.
• Diagnose failures using execution steps, tool requests, tool responses and correlated Gateway traces.
• Work with QA to define persona scenarios, expected outcomes and conversation-level assertions.
• Manage agent configuration changes, deployment versions and environment-specific settings.

Key Responsibilities

• Hands-on experience with Google CXAS agent configuration, tools, deployment and Canvas Preview, demonstrated through practical work.
• Conversational design: intent handling, clarification, context management, handoffs and recovery.
• Strong understanding of LLM instructions, tool schemas, structured responses and grounding answers in returned data.
• REST APIs, JSON, authentication and debugging service integrations.
• Python or TypeScript proficiency sufficient to implement and troubleshoot integration code.
• Voice-assistant concepts, including speech recognition, speech synthesis, interruption handling and latency.
• Experience testing multi-turn conversations and assessing task completion, factual accuracy and failure handling.
• Google Cloud fundamentals, source control and release practices.

Skill Requirements

• Hands-on experience with Google CXAS agent configuration, tools, deployment and Canvas Preview, demonstrated through practical work.
• Conversational design: intent handling, clarification, context management, handoffs and recovery.
• Strong understanding of LLM instructions, tool schemas, structured responses and grounding answers in returned data.
• REST APIs, JSON, authentication and debugging service integrations.
• Python or TypeScript proficiency sufficient to implement and troubleshoot integration code.
• Voice-assistant concepts, including speech recognition, speech synthesis, interruption handling and latency.
• Experience testing multi-turn conversations and assessing task completion, factual accuracy and failure handling.
• Google Cloud fundamentals, source control and release practices.

Other Requirements

• Dialogflow CX experience, supplemented by demonstrated CXAS capability.
• Telecom customer-service journeys, particularly Billing.
• External state-engine integration and multi-agent architecture.
• Automated conversational evaluation and AgentCert integration.

Telephony integration experience if a phone channel is included in scope