About the role
structured by ORIINFRASTRUCTURE ENGINEER INTERN @ DEDALUS LABS MISSION Dedalus Labs is an AI research neolab building infrastructure for AI agents. We’re building the persistent compute layer that powers the next generation of autonomous software.
What you will do
- Work directly alongside Dedalus engineers on real infrastructure, write production-quality code, investigate difficult problems, and explain your decisions.
- Contribute to distributed infrastructure for large-scale AI agent workloads.
- Contribute to persistent compute and distributed storage systems.
- Contribute to scheduling and orchestration platforms.
- Contribute to virtualization, containerization, and sandboxing infrastructure.
- Contribute to reliable multi-tenant cloud systems.
- Contribute to internal developer platforms and infrastructure tooling.
- Contribute to performance, reliability, observability, and failure recovery.
- Work on production systems operating under real-world latency and fault-tolerance constraints.
What they are looking for
- A public GitHub profile is required.
- Demonstrated programming ability in Rust, Go, C, C++, or a similar language.
- Evidence that you have built and completed technically meaningful projects.
- Solid computer-science fundamentals and a genuine interest in systems.
- The ability to reason clearly about tradeoffs, failure modes, and debugging.
- Specificity about what you personally built and why you made particular decisions.
- High agency, intellectual honesty, curiosity, and learning velocity.
- Strong written and verbal technical communication.
- The ability to work full-time and in person from our San Francisco office for approximately three months.
- A specific reason you want to work on infrastructure for AI agents at Dedalus.
Nice to have
- Self-directed systems projects built beyond required coursework.
- Meaningful open-source contributions.
- A storage engine, scheduler, database, runtime, compiler, operating system, hypervisor, container system, or distributed service you built.
- A correct implementation of Raft, Paxos, replication, sharding, or another distributed-systems concept—with thoughtful testing and documentation.
- Technical writing that explains architecture, failure modes, and tradeoffs.
- Evidence that other people use or depend on something you built.
- Research accompanied by a working implementation.
- A project where you encountered a difficult failure and can explain how you diagnosed it.
- Repeated evidence of unusually fast learning and end-to-end ownership.
- Experience with distributed storage systems.
- Familiarity with consistency models, consensus algorithms, or replication protocols.
- Experience with Kubernetes or modern cloud infrastructure.
- Experience with virtualization, hypervisors, containers, or Firecracker.
- Kernel, operating-systems, or low-level runtime experience.
- Experience with concurrency, networking, or performance engineering.
- Contributions to systems-focused open-source projects.
- Published or ongoing systems research.
- Experience operating infrastructure used by real users.
- Familiarity with infrastructure for AI agents or long-running workloads.
Benefits
- Paid internship
- Meals and office benefits included
- Relocation support may be available
- Visa support may be considered
Full posting text
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.
We're looking for someone who notices a slow build, a repeated manual step, or a service that fails in a way nobody can explain, and wants to make it better. You'll improve the team's development tools and workflows through focused fixes and speed improvements.
ABOUT THE INTERNSHIP
This is a paid, full-time internship or co-op in San Francisco lasting 3-9 months, depending on the program and agreed scope. Applications are open throughout the year, including off-cycle terms. Share your available start date and duration. You'll work alongside Dedalus engineers with direct feedback on your implementation and decisions.
We want evidence that you can learn something difficult, finish a working version, and improve it when the first approach fails. You do not need prior production ownership. You should already have a project you understand well enough to explain its hardest bug, its limits, and your personal contribution.
Your main work will be fixing bugs and papercuts, simplifying code, improving tests and tooling, and making focused speed improvements. You will work within existing systems with support from an engineer. This is how you learn the codebase and our standards. Responsibility grows as you demonstrate sound judgment, reliable execution, and clear communication.
WORK YOU MIGHT TAKE ON
- A build, test, or deployment improvement. Trace where the time goes, identify repeated work or a reliability problem, and verify the change against the same workload.
- A provisioning or developer-tooling workflow that another engineer can run and troubleshoot from its documentation.
- A small code change that eliminates repeated work, improves build or cache efficiency, or speeds up a developer workflow, with tests for the behavior it changes.
- Diagnostics that connect a failed operation to the relevant logs, metrics, and resource state, followed by a fix for the underlying problem.
Your project will have an agreed scope and completion criteria. The outcome should be usable by the team, with tests, documentation, and a clear account of its limitations.
WHAT WE LOOK FOR
You have completed a meaningful programming project in Go, Rust, C, C++, or a similar language. You can explain its design, inspect its behavior on Linux, and investigate a failure. Experience with CI/CD, containers, Kubernetes, cloud infrastructure, networking, or infrastructure as code is useful.
Show us a tool someone used, a service you operated, a difficult bug you reproduced, or an experiment that led to a working improvement. Research and coursework are welcome too. Tell us what you personally implemented, including any decisions or extensions beyond the starting instructions.
When you get stuck, you narrow the problem, read the relevant source or documentation, and try an experiment that can teach you something. You ask for help with the evidence you've collected and explain what you need to understand next. You act on feedback and keep ownership of the result.
You use AI coding tools where they help. You must be able to explain the resulting code, test its behavior, and correct it when it is wrong. We care about your judgment and understanding throughout the work.
We evaluate the work relative to your experience and opportunities. A degree, a particular employer, or owning hardware is not a prerequisite. Be specific about what you have already learned and where you want to go deeper.
THIS INTERNSHIP MAY BE A POOR FIT IF
- You want to configure tools without understanding or investigating the software they run.
- You need every next step supplied before trying a small experiment or asking a focused question.
- You declare an improvement without checking whether another engineer can use it and whether the result is measurable.
LOGISTICS
- Rolling applications throughout the year.
- An agreed term of 3-9 months, depending on the co-op program and scope. Share your available dates, expected graduation date, and earliest possible full-time start.
- Paid, full-time, and five days per week in person in San Francisco.
- Meals and office benefits included.
- Relocation support may be available.
- Visa support may be considered depending on the candidate and circumstances.
- Possible consideration for full-time employment, without a guarantee of conversion.
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.
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.
COMPENSATION
Off-cycle internships pay $30-$40 per hour. Summer internships pay $45-$50 per hour. Actual pay is based on hours worked. At 40 hours per week, this is approximately $5,200-$6,933 per month off-cycle or $7,800-$8,667 per month in summer. We agree the applicable rate and dates before the internship begins.