Software Engineering, Compute Infrastructure

About the role

Dedalus Labs builds persistent computers for AI agents. Our flagship product, Dedalus Machines, gives agents an isolated environment where they can run software, keep files and state, and work over time.

Our first GPU offering will give customers SSH access to bare-metal GPU servers. The first job is to bring that capacity online and make it dependable: machines that work, secure access, clear documentation, useful performance, and an operating process customers can rely on.

We're looking for a strong systems engineer to own this work with our CTO. You'll work with datacenters and hardware vendors, write software, debug failures, and stay responsible after customers start using the machines. We'll build the software layer incrementally as customer needs and operating experience show us what is worth building.

You'll be the founding engineer and technical lead of our compute infrastructure team, working directly with the CTO. You'll start hands-on, establish the technical and operating foundations, and help hire and mentor the team as it grows. We value people who can make progress with a small team and simple tools, then improve the system as its requirements become clearer.

What you'll work on first

Bring up and operate bare-metal GPU capacity. Evaluate colocation options and work through vendor terms, capacity, power, cooling, networking, and service-level agreements. Coordinate racking, cabling, remote management, and hardware acceptance. Install and validate Linux, firmware, GPU drivers, storage, and networking so customers can connect over SSH and run their workloads.

Make the service dependable. Own the practical systems administration and site reliability work: user and SSH access, software environments, system services, configuration, patching, monitoring, and incident response. Investigate failures across hardware, firmware, the operating system, and customer workloads. Coordinate repairs and maintenance with vendors. Measure availability and performance, keep customers informed, and turn recurring problems into durable fixes.

Build useful software as the need appears. Automate repeated setup and operating work. Make provisioning, configuration, health checks, upgrades, and recovery repeatable and testable. Reuse existing tools where they fit and choose improvements that make a measurable difference to customers or the team.

Where the work can grow

Compute infrastructure spans many layers. The areas below describe relevant directions for this role. We'll choose which to pursue based on customer demand, operating needs, and your strengths. You can bring depth in one or more areas and learn the layers around them.

• Machine and cluster lifecycle: discovery, inventory, allocation, provisioning, configuration, upgrades, maintenance, recovery, and decommissioning.

• Fleet health and reliability: hardware diagnostics, failure isolation, automated repair, observability, safe rollouts, and reducing unavailable or unusable capacity.

• Systems and performance: Linux, kernels, drivers, firmware, networking, storage, and the way workloads use compute and memory. Use profiles, benchmarks, and production evidence to guide changes.

• Infrastructure software and interfaces: services, controllers, CLIs, and APIs that make compute easier to request, understand, and operate. Reason about durable state, concurrency, permissions, and recovery when operations fail partway through.

• Scheduling and capacity: placement, resource accounting, hardware topology, isolation, and efficient use of the fleet as workloads and customer needs evolve.

• Virtualization and product integration: GPU assignment, guest I/O, device isolation, and integration with Dedalus Machines where the product calls for it.

You'd be a good fit if

• You've built or operated compute infrastructure that other people depend on. You can explain your own contribution, the constraints, what failed, and how you established the result. Hands-on GPU deployment and datacenter experience are especially relevant to our initial work.

• You have strong software engineering fundamentals and depth in Linux, distributed systems, networking, storage, virtualization, or another relevant systems area. You write reliable production code in Go, Rust, C++, Python, or a comparable language and can follow a problem across layers.

• You can turn an incomplete goal into a concrete next step. You make practical choices about what to automate, what to reuse, and what to build later. You take responsibility for correctness and for the people using the system.

• You communicate clearly, work constructively through disagreement, and remain calm during incidents. You can work with datacenter operators, vendors, customers, and engineers, and help others become effective.

• You use modern AI coding tools to accelerate exploration, implementation, testing, and review. You understand the resulting code and take responsibility for its behavior.

Experience we value

We place particular weight on experience operating server fleets at scale and running production Kubernetes. Tell us about the fleet or clusters you owned, the workloads, availability and performance requirements, upgrades, incidents, and the improvements you personally delivered. Experience building Kubernetes controllers or operators and managing node lifecycle, networking, storage, or capacity is especially useful.

Practical Linux systems administration and site reliability engineering experience matter: access and identity, SSH, system services, package and configuration management, patching, observability, incident response, and recovery. We value engineers who understand the operating system and can write software to make these tasks reliable across a fleet.

HPC tooling is a plus. Relevant examples include Slurm scheduling, queues and accounting, Environment Modules (module load), shared software environments, and Python/CUDA dependency management. Experience with Ansible or other configuration-management tools, infrastructure as code, and modern toolchains such as uv is useful. We care about your ability to make environments reproducible and understandable for the people running workloads.

Other relevant depth includes GPU hardware health, Linux systems software, distributed services, high-performance networking, storage, and virtualization such as KVM or QEMU. These are useful backgrounds, and we will choose tools and software capabilities as the product needs them.

We prefer demonstrated success in production. New graduates and early-career candidates are welcome when exceptional relevant work demonstrates the ownership, judgment, and depth this role requires. We evaluate the work and its results, including substantial research, open-source, or independent work. You do not need experience with every possible direction above.

Logistics

• Full-time and in person in San Francisco.

• Travel to datacenters as deployment and operational work requires.

• Relocation support is available.

• We sponsor visas for exceptional talent.

• Annual base salary: $230,000-$350,000 USD, based on experience and role scope.

• Equity is under discussion. We are considering 1%-5%, with the final grant to be agreed.

• Meals and office benefits are included.

Show us your work

Use your application to show us one thing you built and one difficult problem you investigated. They can come from the same project. Choose work you already have. We are not asking you to build a new project for this application.

• Work: Show the code, demo, design, or result. Explain what you personally owned, what you reused, and one decision that made the result better or simpler.

• Investigation: Describe the symptom, your first explanation, the evidence that changed your thinking, and how you established the result. Include the workload or users involved and any important limitations.

• Writing: Share one piece of writing you care about as a separate, required sample. Any medium or length is welcome. Tell us why you wrote it, who it was for, and what you wanted to communicate. Looking back, what works well and what would you change? We care about the quality of your thinking and writing, and your reflection on the work. Link it or paste a redacted excerpt. Explain your contribution and any coauthors or AI assistance.

For this role, a useful example is a GPU cluster you brought online, a provisioning service you built, a virtualization problem you solved, or a production failure you diagnosed and prevented from recurring. Explain your own contribution and what happened after people depended on the system.

Work from employment, open source, research, coursework, or independent projects is welcome. Public source code is optional. A redacted excerpt or a concrete account of private work is fine. Keep confidential information out of your application.

Tell us which problem in this role you want to own and why. If your strongest evidence is easy to miss on a resume, point us to it. An existing demo or short screen recording is welcome, and video is optional.