Job Description

Summary

Own end-to-end development of autonomous capabilities (perception ?

localization ? planning ? control) with a strong emphasis on simulation-first

workflows. Lead ROS 2 architecture, sensor integration, Nav2 customizations, and

motion planning, delivering features that transfer from sim to real robots. Responsibilities

? Simulation-first delivery

o Build/maintain high-fidelity worlds, robots, and sensors in

Gazebo/Ignition and/or Isaac Sim; create reproducible SIL/HIL

pipelines and regression suites. o Develop domain-randomization and sensor-noise models to improve

sim-to-real transfer. ? Sensors & Estimation

o Integrate and calibrate LiDAR, camera, depth, radar, IMU, GNSS;

manage extrinsics and time sync (PTP/NTP). o Implement and tune EKF/UKF (e.g., robot_localization) and map/pose

graph maintenance. ? Navigation & Motion Planning

o Own Nav2 stack: behavior trees, costmaps, planners/controllers;

obstacle layers and recovery behaviors. o Implement global/local planners: sampling-based (RRT*, PRM) and

trajectory optimization (CHOMP, TrajOpt, MPC). ? ROS 2 Architecture

o Design scalable node graphs, lifecycle nodes, parameters, and TF

trees; tune DDS QoS for reliability/latency. o Build robust launch systems, tooling, and observability (rosbag2, tracing, metrics). ? Performance & Re

o Profiling (perf, valgrind, sanitizers), latency budgets, CPU pinning;

PREEMPT_RT (nice-to-have). o Establish test strategy: unit/integration/e2e in sim, automated

scenario testing, fault injection, and KPIs. ? Collaboration & Leadership

o Write clear design docs/RFCs; mentor teammates; partner with

hardware for bring-up and with ML for perception models. Minimum Qualifications

? 6-10+ years in robotics; 3+ years with ROS 2 (Foxy+). ? Strong C++17/20 and Python; solid Linux/Ubuntu. ? Deep hands-on with Gazebo/Ignition (or Isaac Sim) for building

robots/worlds, plugins, and sensor models.

? Proven work with sensor fusion (IMU/LiDAR/camera/GNSS) and calibration

pipelines. ? Nav2 expertise (BT customization, costmaps, planners/controllers) and

SLAM (slam_toolbox/Cartographer). ? Motion planning experience: at least one of sampling-based or trajectory

optimization, plus constraints/collision checking (FCL). ? Fluency with colcon/ament, CMake, Git, CI; debugging with gdb and tracing

with ros2_tracing. Nice-to-Have (NVIDIA Ecosystem & Acceleration)

? Isaac Sim/Omniverse workflows, USD assets, synthetic data pipelines. ? Deployments on Jetson (CUDA, TensorRT, nvblox, Isaac ROS GEMs), ONNX/TensorRT model packaging. ? DDS vendor tuning (Fast DDS, Cyclone DDS); SROS2 (security). ? Real-time (PREEMPT_RT), EtherCAT/CAN basics; micro-ROS/RTOS

exposure. Tooling Stack

? ROS 2 (rclcpp/rclpy, TF2, Nav2, slam_toolbox/Cartographer, RViz, rqt). ? Sim: Gazebo/Ignition, (bonus: Isaac Sim). ? Perception: OpenCV, PCL; (bonus: PyTorch/TensorRT for inference

wrappers). ? Testing/Obs: rosbag2, pytest/rostest, scenario runners, metrics dashboards.

? Build/CI: CMake, colcon, GitHub Actions/Jenkins; Docker for reproducible

dev. What Success Looks Like (first 6-12 months)

? 90%+ of nav regressions automated in sim; green within 20 min per PR. ? Reliable Y% success on benchmark

routes/scenarios. ? Robust sim-to-real transfer:
(tracking error, collision rate, time-to-goal). ? Reduced field bugs via failure injection and scenario coverage. Interview Signals

? Practical: build a Nav2 BT plugin to handle dynamic obstacles; instrument

and reduce planning latency with QoS/threads. ? Design: draw a ROS 2 graph for multi-sensor fusion + Nav2, include TF tree

and lifecycle, discuss failure modes. ? Simulation: create a sensor model with noise/latency and show how you'd

validate it against real logs. ? Depth probes: extrinsics/time-sync strategy; local vs global planner

tradeoffs; MPC vs sampling; ros2_tracing usage.

Short checklist against your original list (all covered) ? Kinematics/Dynamics, Control, Planning/SLAM, Perception ?

? Simulation frameworks & physics engines ? (emphasis on Gazebo/Ignition;

Isaac Sim as plus) ? AI/ML for perception (good to have) ?

? Linux, C++/Python, ROS 2 Core ?

? Navigation2, Sensor Fusion, Calibration ?

? DevEx & Reliability (CI, tracing, bags, KPIs) -- explicitly added for senior

roles

Job Type: Full-time

Pay: ₹1,000,000.00 - ₹2,500,000.00 per year

Work Location: In person