About the Role

Apkudo builds the intelligence layer for the device lifecycle. Machines deployed in customer facilities inspect, test, clean, sort, store, and retire devices at industrial volume, and applications running alongside them carry our inspection and diagnostic models. Device OS is the layer all of it is reached through — one interface to communicate with a machine or an application, configure it, monitor it, and deploy new software to it.

We’re looking for a Lead Software Engineer to build that layer. You will own the integration between the platform and everything running in the field, the mechanism that deploys Apkudo software into the ecosystem, and the remote monitoring that makes a distributed fleet legible. When a new computer vision model is ready, you own the path that gets it onto the fleet safely. When an agent is ready to run in a customer operation, you own how it is deployed, observed, and rolled back.

What You’ll Do

• Build the integration layer. The services and contracts that connect machines and applications in the field to the Apkudo platform — identity, configuration, command, and event flow in both directions.
• Own the deployment mechanism. How Apkudo software reaches the ecosystem: application releases, computer vision and diagnostic models, and agents. Packaging, distribution, staged rollout, versioning, and rollback across a heterogeneous fleet.
• Build remote monitoring. Telemetry, health, and diagnostics for every machine on the platform, and the alerting and investigation tooling built on top of it.
• Deliver a unified control surface. A single place to see, configure, and operate machines and applications, rather than a different path for every machine type.
• Partner with robotics engineers on interface design. You are the counterpart who answers “how should my machine talk to the platform,” and you are accountable for that answer being a good one. Developer experience is part of the job, not an afterthought.
• Partner with design on the operator experience. Monitoring and configuration are an operator-facing product, not an internal console with a table in it.
• Collaborate across the platform. Work with Device AI on getting models and agents from training into production on real hardware, and with Device Passport on the event schema everything in the field emits.

What You’ll Bring

• Significant experience building and shipping production backend or full-stack software, including technical leadership of a substantial system.
• Strong distributed systems fundamentals: services, APIs, messaging, and the failure modes that come with an unreliable network between you and the thing you are managing.
• Experience with deployment or release infrastructure — artifact distribution, versioning, staged rollout, rollback — and a clear point of view on what makes a deploy safe.
• Experience with observability in practice: instrumentation, metrics, logs, alerting, and remote diagnosis of something you cannot physically touch.
• You design interfaces as products. You have written an API or SDK that other engineers depended on, and improved it based on how they actually used it.
• Comfort working across an organizational boundary with people who are not software engineers — hardware, robotics, operations — and translating between what they need and what the platform can offer.
• Willingness to contribute to user-facing surfaces, or to partner closely with design to get them right.
• Genuine engagement with where AI is going — using modern AI tooling in your own work, and interested in the problem of getting models and agents deployed and operating reliably in production.
• Authorized to work in the United States. This role is US-based.

Nice to Have

• Experience with fleet, edge, or IoT device management at scale — provisioning, configuration, over-the-air update, or mobile device management.
• Experience with model deployment or MLOps, particularly getting models onto constrained or edge runtimes.
• Experience building for a heterogeneous fleet you do not fully control, where each unit has a different configuration and cannot simply be replaced.
• Platform or developer-tools work where external teams were the customer, including conformance or certification test suites.
• Hands-on experience deploying agentic systems, or with agent tooling and protocols such as MCP.
• Domain exposure to industrial automation, manufacturing systems, logistics, repair, or recommerce.

What Success Looks Like

• Models, applications, and agents reach the fleet through a deployment path you built, with a rollback the team trusts.
• Operations relies on remote monitoring to answer why a machine is behaving differently today, rather than working around it.
• Teams outside Device OS build against your interfaces without needing you in the room.
• Robotics engineers treat the platform interface as something that helps them ship rather than something they work around.
• New machines and applications come onto the platform faster over time, because the integration path is a known one.