Robotics Perception and Edge AI Engineer

About the Role

CDSPACE is looking for a Robotics Perception and Edge AI Engineer to develop real-time perception and intelligent navigation capabilities for UAV and robotic platforms.

You will build computer-vision and sensor-processing pipelines, integrate perception systems with ROS/ROS2, and deploy optimized AI models on edge-computing platforms. The role spans algorithm development, sensor integration, simulation, onboard deployment, bench testing, and flight validation.

What You’ll Do

Robotics Perception

Develop perception algorithms for object detection, tracking, classification, localization, and scene understanding.

Build real-time image-processing pipelines using OpenCV and GStreamer.

Develop perception capabilities using monocular, stereo, thermal, depth, and other camera systems.

Integrate data from cameras, IMUs, LiDAR, GPS/GNSS, optical-flow sensors, and other UAV payloads.

Support sensor calibration, synchronization, coordinate transformations, and timestamp alignment.

Improve perception performance across changing lighting, motion, altitude, weather, and operating conditions.

Computer Vision and Machine Learning

Develop, train, evaluate, and improve computer-vision and deep-learning models.

Prepare datasets and define data-collection, annotation, cleaning, and augmentation workflows.

Select model architectures based on accuracy, latency, memory, power, and hardware constraints.

Define evaluation metrics and conduct structured model-performance analysis.

Investigate model failures, dataset limitations, false detections, and edge cases.

Maintain reproducible training, validation, and experiment-tracking workflows.

Edge AI Deployment

Deploy AI models on NVIDIA Jetson, Raspberry Pi, and other edge-computing platforms.

Optimize models for real-time inference using quantization, pruning, hardware acceleration, and other optimization techniques.

Convert and deploy models using ONNX, TensorRT, or similar inference frameworks.

Profile CPU, GPU, memory, power consumption, throughput, and end-to-end latency.

Integrate inference pipelines with cameras, sensors, communication systems, and flight software.

Develop reliable startup, monitoring, logging, recovery, and health-check mechanisms for onboard AI applications.

ROS and Robotics Integration

Integrate sensors and perception pipelines with onboard computing platforms using ROS/ROS2.

Develop ROS/ROS2 nodes, services, topics, actions, and launch configurations.

Build and test robotic applications using Gazebo or similar simulation environments.

Integrate perception outputs with guidance, navigation, control, and mission-planning systems.

Develop interfaces between onboard computers, flight controllers, payloads, and ground-control systems.

Diagnose issues across sensors, middleware, networks, embedded software, and onboard computers.

UAV and Autonomous Systems

Develop perception and navigation capabilities for multirotor and fixed-wing UAV platforms.

Support obstacle detection, target tracking, visual navigation, inspection, and autonomous mission workflows.

Integrate onboard perception applications with PX4, ArduPilot, or similar flight stacks.

Support communication with flight controllers using MAVLink.

Work with QGroundControl, Mission Planner, and related UAV tools.

Collaborate with avionics, embedded software, controls, hardware, and flight-test teams.

Testing and Validation

Conduct software-in-the-loop and hardware-in-the-loop simulations.

Develop test datasets, test scenarios, and acceptance criteria for perception systems.

Perform bench testing, field testing, and flight trials.

Evaluate system performance under representative operating and environmental conditions.

Diagnose sensor, communication, model, software, and system-level issues.

Perform root-cause analysis and implement corrective actions.

Maintain testing procedures, integration records, model documentation, and validation reports.

Application and Interface Development

Develop supporting applications, visualization tools, and operator interfaces.

Build tools for model configuration, telemetry monitoring, data recording, playback, and analysis.

Support Qt-based software development when required.

Implement clear logs, diagnostics, and system-health reporting.

Maintain modular, testable, and documented software.

Required Qualifications and Skills

Strong proficiency in Python.

Working proficiency in C/C++ for performance-critical and hardware-integrated applications.

Experience developing and deploying computer-vision or deep-learning models.

Experience with OpenCV and real-time image-processing pipelines.

Familiarity with GStreamer or similar multimedia frameworks.

Experience with a deep-learning framework such as PyTorch or TensorFlow.

Understanding of model training, validation, dataset preparation, augmentation, and performance evaluation.

Experience deploying AI models on edge-computing platforms.

Familiarity with ROS/ROS2 and robotic software architectures.

Experience integrating cameras, sensors, and onboard computing systems.

Understanding of Linux-based development environments, networking, and system debugging.

Ability to profile and optimize inference latency, throughput, memory usage, and power consumption.

Strong analytical, debugging, and problem-solving skills.

Ability to communicate technical findings clearly and work effectively across engineering teams.

Good to Have

Experience with NVIDIA Jetson, CUDA, TensorRT, ONNX, or GPU-accelerated computing.

Experience with Raspberry Pi or other embedded Linux platforms.

Knowledge of object detection, tracking, segmentation, visual odometry, SLAM, or sensor fusion.

Experience with ROS/ROS2, Gazebo, and robotics simulation.

Knowledge of MAVLink, QGroundControl, and Mission Planner.

Hands-on experience with PX4, ArduPilot, or a similar flight stack.

Understanding of UAV flight dynamics and multirotor or fixed-wing architectures.

Experience with Qt software development.

Experience with camera calibration, synchronization, streaming, and recording.

Familiarity with MLOps, experiment tracking, dataset versioning, and automated model evaluation.

Experience taking perception or AI systems from research prototypes through field deployment.

Experience: 3–5 years in computer vision, robotics perception, AI/ML, or edge computing.

Education: Degree in Electronics, Electrical Engineering, Computer Science, Artificial Intelligence, Robotics, or a related engineering discipline.