New opportunity

Practice Customer Engineer, Artificial Intelligence, Google Cloud

Google · Sydney, Australia

About this role

The Practice Customer Engineer will specialize in Gemini enterprise, partnering with technical Sales teams to help customers integrate generative AI into secure, data-connected enterprise solutions on Google Cloud. The role involves understanding customer business and technical requirements, presenting tailored solutions, and demonstrating the value of Google Cloud’s AI portfolio.

Responsibilities

Drive technical adoption of Gemini enterprise workloads, supporting customers from technical evaluation through ramp-up. Recommend integration strategies, enterprise architectures, platforms and application infrastructure for complete Google Cloud solutions.
Combine business strategy, development and prototyping to build functional solutions tailored to customers and gain buy-in from domain experts. Advise customers on enterprise AI integration patterns and build technical relationships.
Demonstrate and prototype integrations in customer and partner environments. Document, prioritize and help resolve customer feature requests and issues with product and engineering teams. Travel to customer sites, conferences and related events as required, representing Google Cloud publicly.

Minimum qualifications:

Bachelor’s degree or equivalent practical experience, and 10 years of experience with cloud-native architecture in a customer-facing or support role, including application development and integration.
Experience with search systems, including retrieval, ranking and search quality tuning; presenting to technical stakeholders and executive leaders; and coding in Python, JavaScript or TypeScript, Go, or Java to demonstrate or prototype integration patterns.
Experience architecting solutions that connect AI models and agents to enterprise data using Retrieval-Augmented Generation (RAG), Text-to-SQL and semantic search.

Preferred qualifications

Experience developing agents with LangGraph, Semantic Kernel or Google AI Agent Development Kit; using iPaaS, API gateways or Enterprise Service Buses in cloud environments; and applying functional metrics to evaluate models and agents.
Knowledge of distributed tracing, logging and audit logging for AI applications; integration governance and security, including short-lived credentials; and OpenAPI and Model Context Protocol integration patterns for connecting agents with business systems and API gateways.

The role is based in Sydney NSW, Australia. Google states that English proficiency is required for its roles unless otherwise specified.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

Google CloudGemini enterpriseGenerative AICloud-native architectureApplication developmentApplication integrationEnterprise architectureSearch systemsRetrievalRankingSearch quality tuningPythonJavaScriptTypeScriptGoJavaAI agentsRetrieval-Augmented Generation (RAG)Text-to-SQLSemantic searchLangGraphSemantic KernelGoogle AI Agent Development Kit (ADK)Integration Platform as a Service (iPaaS)API gatewaysEnterprise Service Buses (ESBs)Model evaluationAgent evaluationDistributed tracingLogging and audit logging

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Drive Gemini enterprise adoption
Partner with technical Sales teams from customer evaluation through ramp-up, recommending the integration strategy and Google Cloud architecture needed for a complete solution.
Architect secure, data-connected AI
Advise on connecting agents and models to enterprise data and business systems. The posting specifically names RAG, Text-to-SQL, semantic search, integration governance and short-lived credentials.
Prove value through working prototypes
Develop and demonstrate customer-tailored integrations in customer or partner environments, combining business requirements with hands-on development to gain domain-expert buy-in.
Establish search and AI quality
Apply retrieval, ranking and search-quality tuning expertise; the posting also prefers functional metrics for assessing model and agent quality.
Translate customer needs across teams
Present to technical stakeholders and executives, build advisory relationships, and document and prioritize feature requests and issues with product and engineering teams.

Likely challenges to think through

Balance speed with enterprise controls
Likely preparation challenge, not a stated employer problem: a persuasive prototype may need to become a secure, governable integration. Prepare to explain architecture choices, credential handling, observability and what would change before broader adoption.
Make AI quality demonstrable
Likely preparation challenge: a fluent response does not establish that retrieval, ranking or an agent is reliable for a customer’s data. Prepare a concrete evaluation approach using search-quality checks and functional metrics, including how you would investigate weak results.
Connect technical detail to business buy-in
Likely preparation challenge: domain experts, technical stakeholders and executives may need different evidence during evaluation and ramp-up. Prepare to explain one integration at each level, including its business purpose, working demonstration, trade-offs and unresolved issues.

Practice your answers

Technical practice

How would you architect a secure, data-connected Gemini enterprise solution on Google Cloud?

What to coverThe role calls for enterprise architecture and application integration, not just a chat interface. Suggested preparation: explain the major components, data flows, security boundaries, and trade-offs you would validate against customer requirements.

How would you choose between RAG, Text-to-SQL, and semantic search for an enterprise use case?

What to coverThe posting names all three patterns for connecting AI to enterprise data. Suggested preparation: compare the data and question types each suits, how answers would be grounded, and where access controls or answer quality could become a concern.

How would you diagnose and improve poor search relevance?

What to coverExperience with retrieval, ranking, and search quality tuning is a minimum qualification. Suggested preparation: describe how you would distinguish retrieval failures from ranking failures, assess relevance, and test a change before rollout.

How would you connect an AI agent to business systems through APIs?

What to coverThe preferred qualifications mention OpenAPI, Model Context Protocol, and API gateways. Suggested preparation: discuss interface contracts, permitted actions, failure handling, and when each integration pattern would be useful.

What would guide your choice of an agent framework?

What to coverThe posting cites LangGraph, Semantic Kernel, and Google AI Agent Development Kit as examples of preferred experience. Suggested preparation: compare how you would assess orchestration needs, integration fit, observability, and maintainability without assuming one framework is required.

How would you govern and secure integrations between agents and enterprise data?

What to coverThe role calls for secure solutions, with preferred knowledge of integration governance and short-lived credentials. Suggested preparation: cover authorization boundaries, credential lifecycle, auditability, and the trade-off between access and operational simplicity.

When would you use an iPaaS, API gateway, or ESB in a cloud integration?

What to coverThe posting lists these as preferred integration experience. Suggested preparation: compare their roles in connecting systems, exposing APIs, and managing integration complexity, including how an existing customer environment would affect the choice.

How would you prototype a customer-specific integration?

What to coverThe role involves coding to demo or workshop integration patterns and prototyping in customer or partner environments. Suggested preparation: outline a small functional prototype in one of the listed languages, the assumptions it would test, and how you would keep its scope distinct from a production design.

How would you evaluate model and agent quality for a customer workload?

What to coverFunctional metrics for model and agent quality are a preferred qualification. Suggested preparation: define measures tied to the customer's task, explain how you would build representative test cases, and discuss quality trade-offs before broader adoption.

How would you troubleshoot an AI application spanning models, agents, and enterprise systems?

What to coverThe posting identifies distributed tracing, logging, and audit logging as relevant observability constructs. Suggested preparation: explain what each would reveal about a failing request, how you would isolate the fault, and how you would handle sensitive information in diagnostic records.

Behavioural practice

Tell me about a time you uncovered a customer's underlying technical and business requirements.

What to coverUnderstanding both kinds of requirements is central to the role. Suggested preparation: use a real situation, explain the questions you asked and how you reconciled competing needs, then state the resulting decision or outcome.

Describe a time a prototype helped you gain buy-in from domain experts.

What to coverThe role calls for functional, customer-tailored prototypes that secure domain-expert buy-in. Suggested preparation: describe the situation, what you built or changed in response to feedback, and the observable result.

Tell me about a time you worked with a sales team while presenting a technical solution to senior stakeholders.

What to coverThe posting calls for partnership with technical Sales teams and presentations to technical stakeholders and executive leaders. Suggested preparation: explain your role, how you adapted the message without losing technical accuracy, and what happened afterward.

Describe a time you prioritized a customer issue or feature request with product or engineering partners.

What to coverThe responsibilities include documenting, prioritizing, and helping resolve customer requests and issues. Suggested preparation: give a real example covering the evidence you gathered, how you handled competing priorities, and the resolution or status you communicated.

Tell me about a time you represented a technical solution in a customer-facing or public setting.

What to coverThe role includes customer consultation and, as required, representing Google Cloud at sites or events. Suggested preparation: describe the audience and situation, how you handled questions or uncertainty, and the outcome; a comparable past setting is sufficient.

Sharpen your knowledge

Topics to prepare

Cloud-native enterprise AI architecture
The role calls for recommending Google Cloud architectures and integration strategies for Gemini enterprise workloads. Revision exercise: sketch a secure, data-connected solution from customer evaluation through ramp-up, and explain the architecture choices and trade-offs.
Search and retrieval quality
Retrieval, ranking and search quality tuning are minimum qualifications. Revision exercise: compare two retrieval-and-ranking approaches for an enterprise document search scenario, identifying relevance failures and how you would measure improvement.
Agents connected to enterprise data
The posting requires experience architecting agent-based AI integrations using patterns such as RAG, Text-to-SQL and semantic search. Revision exercise: choose an appropriate pattern for each of three hypothetical customer questions and explain the data flow and limitations.
Integration security and governance
Preferred knowledge includes short-lived credentials, API gateways, OpenAPI and MCP integration patterns. Revision exercise: draw an agent-to-business-system connection and identify where access control, credential handling and governance decisions belong.
AI evaluation and observability
Functional quality metrics, distributed tracing, logging and audit logging are preferred qualifications. Revision exercise: define what you would measure and record to distinguish a poor answer from a retrieval, agent or integration failure.

Skills to revise

Prototype an integration
Coding customer demos in Python, JavaScript or TypeScript, Go, or Java is a minimum qualification. Practice building a small prototype in one of those languages that connects an agent to a sample enterprise data source, then walk through its integration points.
Tune a search example
The posting explicitly calls for search quality tuning. Practice assembling a small set of sample queries and expected results, adjusting retrieval or ranking, and showing which queries improved or regressed.
Present a customer-tailored solution
The role involves understanding requirements, securing domain-expert buy-in and presenting to technical stakeholders and executives. Practice a short discovery conversation followed by two explanations of the same proposed solution: one architectural and one focused on business value.
Demonstrate an agent integration pattern
As preparation for the preferred agent-framework and OpenAPI/MCP experience, consider a small demonstration using a framework named in the posting to connect an agent to a sample business-system interface. Explain the interface boundaries and security choices; use of a particular framework is not stated as mandatory.
Evaluate and document a prototype issue
The role includes documenting and prioritizing customer issues with product and engineering teams. Practice reproducing one failure in a sample AI integration, recording its user impact and diagnostic evidence, and proposing a priority and next step.

Your preparation checklist

Map evidence to the core requirements
Prepare a one-page inventory of examples covering customer-facing cloud-native architecture, application development and integration, search systems, coding, and presentations to technical and executive audiences. The posting specifies 10 years of relevant experience; use dates and scope only where you can substantiate them.
Build an enterprise AI project walkthrough
Choose a real project, if available, and outline the customer need, architecture, your decisions, prototype, and results. Highlight how agents connected models to enterprise data through RAG, Text-to-SQL, or semantic search; distinguish what you built from what you proposed.
Rehearse a secure integration design
Draw a hypothetical Google Cloud solution for a data-connected Gemini enterprise workload. Explain data access, application infrastructure, security boundaries, and trade-offs; consider API gateways, OpenAPI or MCP, and short-lived credentials as relevant design options, not assumed requirements for every solution.
Prepare a search and evaluation deep dive
Bring an example or practice exercise showing how you would assess retrieval, ranking, and search quality. Define functional measures for model and agent behavior and explain how results would guide tuning.
Make a small integration prototype ready to explain
Prepare code or a clear demo plan in Python, JavaScript or TypeScript, Go, or Java that connects an agent or model to a sample data source or business API. Be ready to explain limitations; an agent framework named in the preferred qualifications is optional.
Practice both customer-facing presentations
Create a brief executive explanation of business value and a technical walkthrough of architecture and trade-offs. Rehearse questions that uncover customer requirements and show how a tailored prototype could earn domain-expert buy-in from evaluation through ramp-up.
Outline how you would handle integration issues
Draft a sample issue record with impact, reproduction steps, logs or traces, priority, and a proposed next action. Use it to explain how you would work with product and engineering teams on customer feature requests or defects.
Prepare questions for the employer
Write focused questions about typical Gemini enterprise evaluations, customer integration constraints, measures of a successful technical win, and how customer feedback reaches product and engineering teams.

Availability last checked: 1 Oct 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in Australia. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…