Research Engineer-ADAS

As a Software Engineer in the field of safety first video telematics products, you will be responsible for developing efficient C++ software for edge devices inside commercial vehicles. Your key responsibilities will include: - Owning C++ software modules for on-device video capture, preprocessing, inference, and post-processing on Linux. - Implementing classical image processing pipelines and computer vision algorithms such as denoise, resize, keypoints, homography, optical flow, tracking. - Building and optimizing distance/spacing estimation from camera feeds using calibration, geometry, and depth estimation networks. - Integrating ML models for driver monitoring and ADAS events like drowsiness, distraction/gaze, phone usage, smoking, seat belt, etc. - Achieving real-time targets in terms of frames per second, latency, and memory on CPU/GPU/NPU using various optimization techniques. - Writing clean and testable C++ code, setting up CMake builds, and following Git-based workflows for version control. - Instrumenting logging/telemetry, debugging and profiling with tools like gdb, perf, valgrind. - Collaborating with data/ML teams on dataset curation, labeling specs, training/evaluation, and model handoff. - Working with product and compliance teams to ensure on-road reliability, privacy, and regulatory compliance. Qualifications required for this role include: - B.Tech/B.E. in CS/EE/ECE or equivalent practical experience. - 2-3 years of experience in computer vision/machine learning or video-centric software roles. - Hands-on experience in modern C++ development on Linux, with proficiency in Git and CMake. - Strong foundations in image processing and computer vision concepts. - Practical experience integrating CV/ML models on devices using tools like OpenCV, ONNX Runtime, TensorRT, etc. - Competence in building real-time video pipelines, optimizing for latency and memory. - Comfort with multithreading, concurrency, debugging, and profiling on Linux targets. As a Software Engineer in this role, you will report to the Technical Lead ADAS. Additional requisites include experience with driver monitoring or ADAS features, knowledge of depth estimation, stereo matching, and familiarity with model training using PyTorch/TensorFlow. Technical competency in languages like C++ and Python, along with proficiency in various CV/ML and video-related tools, is essential. Familiarity with build/DevOps tools and debugging/performance analysis tools is also required for this position.,