Job description

Role Summary: We are seeking an AI Systems Engineer to build the reliable software infrastructure that connects models, applications, external tools, and real-world systems. This role is for someone who enjoys turning AI prototypes into reliable, observable, and maintainable systems. Location: Hollywood, FL (or willing to relocate) What You'll Do: • Design runtimes and services around remote, streaming, or action-sequence model inference. • Handle latency variation, stale outputs, retries, partial failures, service outages, and safe shutdown. • Build interfaces that support different model providers, architectures, data sources, and downstream systems. • Integrate inference clients, state and sensor inputs, task scheduling, observability, and safety or authorization checks. • Build agentic loops and tool-using systems with clear permission boundaries, structured outputs, human approval paths, and MCP-compatible interfaces. • Reproduce failures using deterministic tests, logs, traces, recorded-data replay, and fault injection. • Take prototypes through reliable, observable, and maintainable deployment. Must-Haves: • Strong software architecture, concurrency, networking, debugging, testing, and production-reliability skills. • Experience with agentic systems, tool-use loops, MCP practices, distributed systems, inference infrastructure, or real-time applications. • Programming language flexibility (Python experience is helpful but not strictly required). Nice-to-Haves: • Robotics experience (integration with cameras, robot state, controllers, action scheduling, simulation, or physical hardware). Note: This is a preferred signal rather than a hard requirement. • Experience with ROS or ROS 2, Gazebo, MoveIt, URDF, Xacro, TF2, RViz, rosbag, MCAP, or related robotics tooling.

Requirements

Must-haves

Location: Hollywood, FL (or willing to relocate)
Design runtimes and services around remote, streaming, or action-sequence model inference.
Handle latency variation, stale outputs, retries, partial failures, service outages, and safe shutdown.
Build interfaces that support different model providers, architectures, data sources, and downstream systems.
Integrate inference clients, state and sensor inputs, task scheduling, observability, and safety or authorization checks.
Build agentic loops and tool-using systems with clear permission boundaries, structured outputs, human approval paths, and MCP-compatible interfaces.
Reproduce failures using deterministic tests, logs, traces, recorded-data replay, and fault injection.
Take prototypes through reliable, observable, and maintainable deployment.
Strong software architecture, concurrency, networking, debugging, testing, and production-reliability skills.
Experience with agentic systems, tool-use loops, MCP practices, distributed systems, inference infrastructure, or real-time applications.
Programming language flexibility (Python experience is helpful but not strictly required).

Nice-to-haves

Robotics experience (integration with cameras, robot state, controllers, action scheduling, simulation, or physical hardware). Note: This is a preferred signal rather than a hard requirement.
Experience with ROS or ROS 2, Gazebo, MoveIt, URDF, Xacro, TF2, RViz, rosbag, MCAP, or related robotics tooling.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aipython

Platform: dover
Effort: medium