19 Nov
|
Newspace
|
Bengaluru
19 Nov
Newspace
Bengaluru
- Role Summary
We are looking for a hands-on Computer Vision Engineer to develop real-time perception systems for robotic and aerial platforms. The role combines image processing, visual tracking, camera geometry, and estimation, with a strong focus on deploying reliable software on embedded hardware.
You will own features from algorithm development through integration, performance optimization, and field validation. The ideal candidate has strong computer vision fundamentals, understands robotics and estimation.
Experience: Typically 2-4 years of relevant experience, with demonstrated ownership of deployed systems.
- What You’ll Do
Design and implement real-time image processing, object detection, feature matching and tracking pipelines for EO and thermal cameras.
Develop camera calibration, distortion correction, coordinate transformations, and projection of 3D information onto image frames.
Handle perception challenges involving camera motion, optical zoom, occlusion, changing illumination, and noisy measurements.
Apply estimation and filtering techniques to improve tracking stability and estimate quantities such as object motion and line of sight.
Integrate perception software with robotic systems, including gimbals, vehicle telemetry, and onboard computing.
Build and optimize camera capture, processing, and video streaming pipelines for low latency and efficient CPU, GPU, and memory usage.
Integrate and troubleshoot communication between cameras, onboard computers, and ground systems over Ethernet and wireless networks.
Develop maintainable C++ software on Linux, with attention to concurrency, timestamps, synchronization, and resource management.
Diagnose failures using recorded datasets, logs, profiling tools, and field experiments.
Take ownership of technical design, implementation, testing, documentation, and deployment; contribute to code reviews and engineering decisions.
- Must-Have Qualifications
Strong practical experience in C++ and OpenCV, with Python proficiency for prototyping, analysis, and tooling.
Solid foundations in image processing and computer vision, including feature extraction, tracking, camera models, and geometric transformations.
Working knowledge of linear algebra, coordinate frames, 3D rotations, and rigid-body transformations.
Practical understanding of estimation techniques such as Kalman filtering, measurement uncertainty, outlier rejection, and temporal filtering.
Experience developing and debugging real-time software on Linux.
Understanding of multithreading, buffering, data ownership, and latency in image processing pipelines.
Experience integrating cameras or other sensors into a working perception or robotics system.
Ability to evaluate algorithms quantitatively and investigate failures using experimental evidence.
Bachelor’s or Master’s degree in Electrical, Electronics, Computer Science, Robotics or a related field, or equivalent practical experience.
4.
Strongly
Preferred
Experience deploying computer vision workloads on NVIDIA Jetson or similar embedded GPU platforms.
Experience with ROS / ROS 2, sensor messages, transforms, timestamp synchronization, and recorded-data workflows.
Experience with GStreamer, hardware video encoding/decoding, and effective image transfer between processing stages.
Familiarity with CUDA, TensorRT, model deployment, and inference optimization.
Experience with visual tracking from moving or gimballed cameras, thermal imagery, or cameras with variable zoom.
Knowledge of sensor fusion, visual odometry, pose estimation, or bearing-based estimation.
Familiarity with UAV systems, MAVLink, ArduPilot, or PX4.
Understanding of video compression and streaming over constrained networks using protocols such as RTP, RTSP, or SRT.
Experience validating robotic systems in outdoor environments and turning field observations into measurable improvements.
📌 Engineer III - Robotics (Bengaluru)
🏢 Newspace
📍 Bengaluru