Forward Deployed Agentic Engineer

We are looking for a Forward Deployed Agentic Engineer to design, build, and deploy production-grade agentic AI systems for enterprise customers across New Zealand, Australia, the United States, and the United Kingdom. This is a highly technical, customer-facing role at the intersection of software engineering, AI architecture, product development, and consulting. You will work directly with customers to understand complex business problems and translate them into secure, reliable, and scalable agentic solutions. You will build autonomous and multi-agent systems using technologies such as LangGraph, integrating foundation models, enterprise platforms, data sources, APIs, and business applications. The ideal candidate is an excellent engineer and communicator who can move confidently between customer workshops, solution architecture, hands-on development, production deployment, and ongoing optimisation. Customer discovery and delivery Work directly with enterprise customers to understand their business processes, technical environments, constraints, and desired outcomes Lead technical discovery sessions, architecture workshops, solution reviews, and implementation planning Translate ambiguous business problems into clear technical requirements and executable delivery plans Build trusted relationships across engineering, product, operations, security, data, and executive teams Communicate technical concepts clearly to technical and non-technical audiences Own customer implementations from initial discovery through production deployment and continuous improvement Agentic AI engineering Design, build, and deploy autonomous and multi-agent AI systems using LangGraph and related orchestration technologies Develop workflows involving planning, reasoning, memory, retrieval, tool calling, delegation, approvals, and human-in-the-loop controls Design stateful agent graphs, reasoning loops, routing logic, checkpoints, recovery paths, and long-running workflows Implement persistent and short-term memory strategies for agentic applications Build reusable agent components, tools, nodes, subgraphs, templates, and orchestration patterns Design guardrails that prevent unsafe, unauthorised, or unreliable agent behaviour Debug and improve agent performance using traces, logs, evaluations, test cases, simulations, and production feedback Build model-agnostic systems that operate across multiple foundation-model providers Integrate commercial and open-source models based on security, performance, latency, and cost requirements Work with OpenAI, Anthropic, Google, Azure OpenAI, AWS Bedrock, and self-hosted models Design intelligent model-routing, fallback, retry, and failover strategies Evaluate models for reasoning, classification, extraction, generation, tool use, vision, and structured output Optimise model usage for quality, latency, reliability, context-window requirements, and cost Reduce dependence on any single provider through well-designed abstraction layers Enterprise integrations Connect agentic systems to customer APIs, databases, applications, knowledge sources, and internal platforms Build integrations with CRM, ERP, document management, collaboration, support, productivity, and workflow systems Integrate platforms such as Salesforce, Microsoft 365, Google Workspace, SharePoint, Slack, Teams, ServiceNow, Jira, Confluence, and HubSpot Build secure connectors for databases, data warehouses, vector databases, file systems, object storage, and knowledge bases Implement REST, GraphQL, webhook, SDK, event-stream, and message-queue integrations Design tool-calling interfaces that let agents safely read data, take actions, and interact with external systems Implement OAuth, API keys, service accounts, role-based access controls, and delegated permissions Support emerging interoperability standards and tool-integration approaches where appropriate Retrieval, knowledge, and data Design retrieval-augmented generation and enterprise knowledge systems Build ingestion, chunking, indexing, retrieval, reranking, and citation pipelines Connect agents to structured and unstructured customer data Implement semantic search, metadata filtering, hybrid retrieval, and contextual grounding Handle customer information in accordance with privacy, retention, residency, and governance requirements Improve output accuracy and traceability through grounding, evidence collection, and source attribution Production deployment and reliability Take agentic solutions from prototype through production deployment Build secure, scalable, observable, and maintainable AI applications Deploy using cloud, containerised, serverless, or customer-hosted infrastructure Implement monitoring, logging, distributed tracing, alerting, checkpointing, and audit trails Design recovery from model failures, integration failures, invalid outputs, timeouts, and partial workflow execution Implement retries, fallbacks, circuit breakers, escalation paths, and human approval steps Diagnose complex production issues across model, application, integration, data, and infrastructure layers Work with customer security teams on architecture reviews, threat modelling, access reviews, and deployment approvals Evaluation and agent quality Develop evaluation frameworks for agent behaviour, output quality, tool use, accuracy, safety, latency, and cost Create automated and human-reviewed test suites for agentic workflows Test agents against edge cases, adversarial inputs, ambiguous instructions, and failure scenarios Measure end-to-end business outcomes rather than relying only on model-level metrics Establish regression testing as prompts, models, tools, and integrations change Implement observability and evaluation processes that support continuous improvement Product and engineering collaboration Work closely with product, platform, research, engineering, and customer-success teams Convert recurring customer needs and delivery challenges into reusable product capabilities Contribute to internal frameworks, integration libraries, reference architectures, templates, and developer tooling Provide structured feedback that influences the product roadmap Document architecture decisions, implementation patterns, deployment approaches, and customer learnings Share best practices and help improve engineering and delivery standards What we are looking for Hands-on experience building agentic AI systems with LangGraph Strong understanding of graph-based orchestration, state management, checkpointing, tool calling, memory, routing, and human-in-the-loop workflows Experience with multiple foundation-model providers and model APIs Experience integrating enterprise systems, SaaS applications, APIs, databases, and knowledge platforms Strong knowledge of system architecture, distributed systems, data flows, authentication, security, and observability Experience building production APIs, backend services, integrations, or cloud applications Ability to debug complex issues across models, agent logic, application code, external tools, and infrastructure Excellent customer-facing communication, workshop facilitation, and stakeholder-management skills Strong ownership and the ability to operate effectively in fast-moving, ambiguous environments Willingness to work across international time zones and travel periodically Preferred experience Forward deployed engineering, solutions architecture, technical consulting, implementation engineering, or customer engineering LangChain, LangSmith, or comparable agent-development and observability platforms Commercial and open-source model ecosystems Vector databases, embedding models, semantic search, and retrieval-augmented generation Docker, Kubernetes, serverless architectures, and cloud platforms CI/CD, infrastructure as code, automated testing, and production observability Agent evaluation, prompt evaluation, red teaming, guardrails, and responsible AI practices Regulated industries, enterprise security reviews, privacy requirements, or data-residency constraints Customers across multiple countries, cultures, and time zones A startup, scale-up, consulting firm, or high-growth technology company What success looks like Within your first six months, you will: Build strong relationships with customer engineering, business, and executive stakeholders Deliver production-ready LangGraph-based agentic solutions that solve meaningful business problems Integrate agents with customer systems, model providers, tools, and data sources Establish reliable evaluation, monitoring, security, and operational practices Reduce the time required to move customer use cases from discovery to production Develop reusable integration and orchestration patterns that accelerate future implementations Identify customer requirements that improve the core platform and product roadmap Become a trusted technical advisor to customers and internal teams Working style Enjoys working directly with customers and solving real operational problems Is equally comfortable writing code, designing architecture, facilitating workshops, and explaining technical decisions Can rapidly learn unfamiliar customer systems and industries Takes ownership from initial problem definition through deployment and operational support Prefers practical, measurable customer outcomes over purely experimental AI demonstrations Communicates openly, manages expectations clearly, and raises risks early Works effectively with globally distributed teams Is comfortable when the problem, requirements, and solution are not yet fully defined Location and travel Wellington is preferred, with other locations expected to open in the future. The role supports enterprise customers across New Zealand, Australia, the United States, and the United Kingdom. You should be comfortable collaborating across international time zones, with reasonable flexibility for customer workshops, deployments, and production support. Periodic domestic and international travel may be required. Ready to build production-grade agentic systems with enterprise customers? #J-18808-Ljbffr