Robotics Software Engineer II

Agility's commercially deployed humanoids operate alongside teams in warehouses, manufacturing facilities, and distribution centers-tackling physically demanding and repetitive tasks while enabling workers to focus on higher-value work. With industry-leading safety standards and years of proven deployment data, we're pioneering a new era of automation that enhances human potential. About the Role We are looking for a hands-on Robotics Software Engineer to join our Skills team, focused on developing, testing, validating, and integrating robot skills into real customer workflows.

In this role, you will operate at the boundary between prototype and production, working directly with our humanoid robot to evaluate performance, root-cause failures, and collect targeted teleop data to improve learned behaviors. You will be hands-on with hardware nearly every day, helping bring new customer workflows online and ensuring existing deployments remain reliable. You will define test plans, execute experiments, and validate that our behaviors meet autonomy and performance requirements.

You'll also partner with AI teams to refine learned policies using real-world data.

Key Responsibilities Develop Behaviors for Humanoid Robots & Integrate into Workflows
* Develop and maintain the core library of production-ready skills that form the foundation of customer workflows.
* Script new skills using behavior trees, state machines, or task planners, defining control flow logic and behavior sequencing.
* Deploy classical as well as modern robotics paradigms to implement robot "skills" in order to complete variety of tasks in target domains
* Build and maintain workflows using the robot's skill API.
* Prototype new skills and customer use cases to assess their feasibility. Enable Human-in-the-loop Interventions
* Develop the components needed to allow robots to recover from anomalous situations with human oversight
* Build the data collection pipeline to analyze and catalog task errors, and improve the robot "skills" toolset to resolve these task errors autonomously.
* Create the infrastructure needed to minimize the frequency and duration of interventions where a human is in the loop while maximizing the reach of a single human "resolver" across the humanoid fleet. Testing, Validation & Deployment
* Define what "production-ready" means for each skill or workflow and design the test plans and infrastructure to validate it.
* Recreate corner cases from logs and failures into repeatable tests.
* Validate new skills and workflows both in simulation and on hardware before release.
* Travel occasionally (