Torc Robotics - ML Engineer, II - Road & Lane
Upload My Resume
Drop here or click to browse · Tap to choose · PDF, DOCX, DOC, RTF, TXT
Requirements
• Bachelor’s degree in Computer Science, Robotics, Electrical Engineering, Machine Learning, or a related field with 4+ years of experience, or a Master’s with 2+ years. • Hands‑on experience developing ML models for perception tasks such as lane detection, road surface modeling, multi‑camera fusion, or related geometry estimation. • Strong understanding of camera calibration, multi‑sensor alignment, and projection between image and BEV spaces. • Proficiency in Python and PyTorch, with experience writing production‑quality machine learning code. • Experience training models on large datasets and using scalable compute environments. • Understanding of relevant ML architectures, such as CNNs, transformers, and BEV‑focused perception networks. • Ability to analyze model performance metrics, debug failure cases, and iterate effectively. • Ability to work cross‑functionally with autonomy, perception, and software engineering teams. • Experience working specifically on lane perception, BEV networks, or road topology estimation. • Experience with CUDA kernels or custom PyTorch operations. • Familiarity with SD maps, localization pipelines, or map‑based priors. • Experience with distributed training or large‑scale experimentation frameworks (e.g., Ray). • Publications in major ML/CV conferences (CVPR, ICCV, NeurIPS).
Responsibilities
• Develop and train computer vision and deep learning models for road‑lane detection using monocular and multimodal sensor data (camera, LiDAR, radar). • Build 3D road surface and lane geometry models in BEV space and integrate them into Torc’s autonomy pipeline. • Analyze model performance, identify corner cases, and improve robustness under diverse environmental and long‑tail conditions. • Develop and optimize large‑scale data processing workflows, including annotation, pseudo‑labeling, and data augmentation. • Implement scalable training and evaluation pipelines for lane perception models. • Own deployment-focused work to optimize models for real‑time execution on automotive‑grade hardware. • Leverage SD and HD map priors to improve lane estimation accuracy and stability. • Contribute to architectural discussions, model reviews, and system‑level integration efforts.
Benefits
• $153,200 - $183,300 USD
No credit card. Takes 10 seconds.