# Forward deployed engineerFull-timeRemote (Singapore, Malaysia, Thailand, Vietnam, Indonesia, Philippines)You will design, build, and own the agents and workflows that run inside real client environments: multi-agent pipelines, LLM orchestration, automated workflows, guardrails, evals. You ship production code. You sit close to the problem. Not research. Not slides. Deployed.Apply for this role## About greydigigreydigi is an AI-native orchestration company. We deploy production-ready agentic AI workflows for enterprise and SME clients across Financial Services, Professional Services, and Experience Industries. Our proprietary aironauts framework turns complex business processes into autonomous, agent-driven operations that work in the messy middle of real organisations.## The roleYou will design, build, and ship the agents that power aironauts engagements. Two surfaces: the aironauts framework itself (orchestration, context, guardrails, evals) and the client agents that run on top of it. You write production code. You set the bar for how we build. You own what you ship into client environments.This is not a research role. We deploy.## What you will doFramework development: Extend the aironauts orchestration layer. Build reusable components for model orchestration, context engineering, guardrails, memory, and tool-use interfaces. Decide what becomes platform and what stays project-specific.Agent build: Design and implement multi-agent workflows for client deployments. From scoping with the delivery manager through go-live. Hands on keys, not in slides.Integration: Connect agents to ERP systems, internal APIs, data sources, and downstream actions. Build the plumbing that makes agentic workflows ERP-safe and audit-ready.Context and retrieval: Engineer retrieval pipelines, vector stores, and context layers that ground agents in proprietary client data without leaking it.Guardrails and evals: Build evaluation harnesses, observability, and policy enforcement that let us deploy into regulated environments with confidence.Production readiness: Own deployment, monitoring, and iteration. When an agent fails in production, you fix it. When it works, you make it faster, cheaper, and more reliable.Continuous improvement: Feed engineering learnings back into the aironauts framework. Every client engagement makes the platform sharper.## What you bring* 5+ years of software engineering, including 2+ years building production AI or agentic systems* Deep, current hands-on experience with LLM orchestration frameworks, agent frameworks, and multi-agent system design* Strong working knowledge of retrieval systems, vector databases, and context engineering* Experience with tool-use protocols, function calling, and MCP-style integration patterns* Solid software engineering fundamentals: Python and/or TypeScript, version control, testing, CI/CD* Cloud deployment experience, ideally containerised workloads on a major cloud provider* Familiarity with evaluation, observability, and guardrails for LLM-based systems* Comfortable integrating with enterprise systems: ERPs, internal APIs, identity, audit trails* Fluent English, written and spoken. Clear, direct, no fluff* Active user of AI tools in your own work. Not curious, actually building with them* Based anywhere in Southeast Asia## The offerFull-time, 80 to 100% remote. Competitive salary for your market. Direct access to the founding team. You will shape the technical core of aironauts and the agents that run on it.### Who should not applyThose who prefer prototypes over production, optimise for novelty over reliability, or treat agentic AI as a demo rather than a system real businesses depend on.## The process1. 1 Intro call, 30 min2. 2 Technical deep dive: architecture discussion and live problem walkthrough3. 3 Paid technical exercise, small and time-boxed4. 4 Final conversation with founders5. 5 OfferApply## Apply for forward deployed engineerSend your application, CV included, through our hiring page.Apply for this rolePrefer email? Send your CV to apply+5gzqtrw2uydx9jkw.own-website@inbound.aironauts.ai