Senior Agentic Planning Engineer

Lightspeed•Northbrook, IL

About The Position

As Senior Agentic Planning Engineer, you will own how LightSpeed's robots decide what to do next. You will build a planner that reasons over goals, sequences move and actuate primitives, chooses classical control or a learned policy for each step, and recovers when a step fails. The planner will make decisions continuously as the robot works, using live sensor data and action results, with the goal of enabling robots to complete whole panels or rooms autonomously. You will define primitive and planner-executor contracts; build task graphs from BIM; develop a two-layer planner that combines open-ended reasoning with reach, collision, and trajectory solvers; and determine when to use reasoning models, classical control, or learned policies. You will also develop execution monitoring, recovery, decision logging, safety guardrails, and operator tools for escalations and remote assistance. You will design for future multi-robot coordination, co-own foundation interfaces with world sensing and robot learning principals, and run the new stack alongside production on existing robots. You will prove the system through end-to-end tasks on real robots, set planning and autonomy architecture across both robot families, mentor engineers, and remain hands-on with production code and weekly robot testing.

Requirements

At least 5 years building autonomy, task planning, or decision-making systems for robots or autonomous vehicles, with systems deployed in production.
Experience with task-level planning on real robots, such as behavior trees, task and motion planning (TAMP), HTN, or similar.
Experience with execution monitoring and fault recovery in a deployed system.
Experience building interfaces between planning, perception, and control that other teams used.
Hands-on experience with reasoning models or LLMs in a decision loop, and a clear understanding of where they help and where they do not.
Strong Python and C++ skills.
Experience with ROS2, Linux, Docker, and CI/CD workflows.
A record of setting technical direction across teams and mentoring senior engineers.

Nice To Haves

MS or PhD in Robotics, Computer Science, AI, or a related field.
Behavior or decision planning experience on an autonomous vehicle team.
Experience with LLM or VLM agents for robotics, including constrained action spaces and guardrails.
Motion planning and trajectory optimization experience with MoveIt, OMPL, or similar.
Multi-robot coordination, task allocation, or fleet orchestration experience.
Functional safety or formal verification experience for autonomous systems.
Experience planning from CAD or BIM models.
Background in construction, manufacturing, or industrial automation.

Responsibilities

Define primitive contracts, including inputs, preconditions, success checks, and failure codes.
Define the interface between the planner and executor, including step requests, executor selection, results, and evidence.
Keep the primitive library callable by the planner rather than manually wiring it for each task.
Build task graphs from BIM elements and their dependencies.
Build a two-layer planner with a reasoning layer and solvers for reach, collision, and trajectory.
Determine when reasoning models should be used and keep them off the real-time control path.
Set rules for choosing classical control or a learned policy for each step, based on world-model confidence and step structure.
Extend planner reasoning and sequencing for new trades, and measure how quickly a new trade can be taught.
Build an execution monitor that verifies each step and can retry, switch executors, skip and flag, or escalate to a person.
Log every decision with its reason and evidence so outcomes can be explained.
Keep the planner within guardrails, including a constrained action space, declared parameter ranges, and safety boundaries.
Build the operator view for escalations and remote assistance.
Design the planner so two robots can share a world model and divide work.
Co-own foundation interfaces with world sensing and robot learning principals, and own the primitive contract.
Run the new stack alongside production on existing robots, switchable per job, and test it on real work from the start.
Prove the system by running whole tasks end to end on real robots, including recovery.
Set planning and autonomy architecture across both robot families.
Mentor autonomy and motion planning engineers.
Build a system that engineers can use to teach new trades without the engineer in this role being involved.
Write production code and test on real robots every week.