CoreWeave is The Essential Cloud for AI™. Built for pioneers by pioneers, CoreWeave delivers a platform of technology, tools, and teams that enables innovators to build and scale AI with confidence. Trusted by leading AI labs, startups, and global enterprises, CoreWeave combines superior infrastructure performance with deep technical expertise to accelerate breakthroughs and turn compute into capability. Founded in 2017, CoreWeave became a publicly traded company (Nasdaq: CRWV) in March 2025. Learn more at www.coreweave.com.

We're proud to be a Living Wage accredited Employer.

About the Team

Launchables is CoreWeave's platform for building Physical Intelligence. We bring simulation, training and evaluation tools into a common platform, allowing robotics, autonomous systems and engineering teams to focus on their applications while Launchables handles common execution and infrastructure needs.

We turn specialist Physical AI software into capabilities that researchers and engineers can use, adapt and build upon. Our work combines strong software engineering with applied domain knowledge to make complex workloads accessible, reproducible and useful in real-world environments.

As a Staff Software Engineer, Workloads - Physical AI, you will develop engine integrations, application code, adapters, templates and developer tools that make specialist software available through Launchables. You will work directly with researchers and engineers to understand their tasks, assumptions, dependencies and expected outcomes, translating these into interfaces and configurations that preserve the capabilities they need.

You will take ownership of how users configure, run, customise and evaluate workloads. This includes defining meaningful inputs, defaults, constraints and outputs; building command-line and API interfaces; and creating examples and documentation that help users extend existing templates or author their own. You will establish appropriate validation for packaging, execution and application outcomes, working with domain specialists to make evaluation criteria and their limitations clear.

You will profile workloads to improve preparation time, GPU utilisation, concurrency and data movement while preserving the intended computation. You will also capture the inputs, versions, settings and execution conditions needed to reproduce runs, with logs, metrics and artefacts that help users diagnose failures, compare experiments and interpret results.

Working alongside engineers responsible for the shared runtime and platform services, you will turn feedback from early users into better interfaces, defaults, tests and guidance. You will identify reusable patterns across integrations, shape shared platform capabilities, review designs and mentor colleagues. You will use coding agents and automation to accelerate development, applying the review, testing and domain judgement needed to ensure the resulting software behaves as intended.

Qualifications

Strong software engineering skills, demonstrated through dependable software or developer tools used by other people.
Applied depth in at least one relevant area, such as robotics simulation, autonomous-vehicle evaluation, reinforcement learning, synthetic-data generation or engineering simulation.
Experience making specialist or research software usable beyond its original author, including configuration, testing, documentation and diagnosis.
Sound judgement about experimental assumptions, evaluation criteria, reproducibility and the conclusions a result supports.
Familiarity with containerised execution and experience tracing problems across application and infrastructure boundaries.
Comfortable learning unfamiliar tools and domains, working closely with specialists and recognising when their expertise is needed.
Evidence of improving shared tools or engineering practices across multiple projects, with clear reasoning about what should be common and what should remain application-specific.

Preferred Skills

Experience developing software in Python, Go or C++.