Note: The job is a remote job and is open to candidates in USA. Wand AI builds agentic labor infrastructure that enables humans and AI agents to operate together as a unified workforce. The Staff Software Engineer will architect and build distributed backend systems for agent orchestration, workflow execution, task management, tool calling, sandboxing, and resilient data pipelines at scale.
Responsibilities
- Architect the agent orchestration and workflow execution system on Temporal, including the sandboxing that keeps agent code execution, web access, and file systems locked down and safe
- Build the task management, autonomous execution loops, and tool calling infrastructure that let agents run independently and act in the real world at scale
- Own the distributed data and event pipelines (Kafka, Kubernetes, Redis) that keep the platform observable and resilient
- Turn ambiguous, high stakes problems into concrete technical designs, ship them, and document the architecture so the next engineer doesn't have to guess
- Pair with the team's Tech Lead to raise the bar on architecture and code quality, reviewing designs across the team as it scales
- Work with the Workforce Management team so the execution layer supports what they're building on top, and pitch in on production support, including coverage that overlaps with US afternoons
Skills
- Real scalable distributed systems experience: production systems that run across many machines, not one box
- Hands on with Temporal or comparable workflow orchestration systems
- Strong with Kafka, and Kubernetes for container orchestration (especially for the sandboxing components)
- Deep backend chops. There's no frontend here, so this needs to run deep
- Experience with distributed streaming and event driven systems (we use Kafka and Redis). What you built with these tools matters more than which specific ones you've used
- A track record of owning architecture end to end, not just contributing to someone else's plan
- Strong in Go and Python
- Clear communicator who can design independently and bring the team along
- Exposure to agentic or LLM based systems, especially tool calling
- Experience building safe, isolated execution environments (sandboxing)
- Background in process automation or RPA (e.g. BPMN, Camunda, UiPath, or similar), even without direct agent experience. We're building orchestration for a workforce of agents, and process automation experience transfers well
- Exposure to distributed data stores such as Cassandra or ClickHouse. Helpful, not required
- Experience building infrastructure other teams can't live without
- On call experience for distributed systems
- Time spent scaling infrastructure inside a fast growing engineering org
Company Overview