Sr Machine Learning Engineer

<p><span style="text-decoration: underline;"><strong>Title and Location</strong>:</span> <strong>Sr Machine Learning Engineer in Santa Clara, CA. </strong></p> <p><span style="text-decoration: underline;"><strong>Job

Responsibilities</strong></span></p> <ul> <li>Implement deep-learning models and frameworks for scene segmentation, depth estimation, active learning, and the development of situational awareness for autonomous vehicles operating in construction and agriculture environments.</li> <li>Build and evaluate machine-learning models using data from multiple sensor modalities, including vision, radar, and thermal cameras, and assess trade-offs in accuracy, robustness, and latency.</li> <li>Research and develop new methods to improve detection performance and increase processing speed.</li> <li>Collaborate with robotics engineers to transition algorithms from desktop and server-class systems to real-time field-deployed robotic platforms.</li> <li>Work with systems and software engineers to design and maintain data-processing and annotation pipelines that support continuous system monitoring, evaluation, and improvement.</li> </ul> <p><span style="text-decoration: underline;"><strong>

Qualifications</strong></span></p> <ul> <li>Bachelor’s degree in Computer Science, Electrical Engineering, Computer Engineering, or related field plus 1 year and 6 months of related experience.</li> <li>Required skills: <ul> <li>Train and optimize computer vision models for object detection, semantic segmentation, and monocular/stereo depth estimation using supervised and self-supervised learning, including loss function customization and multi-scale model training (1 yr, 6 mos).</li> <li>Build and evaluate deep learning models using PyTorch and TensorFlow, implementing transfer learning, custom training loops, distributed training, gradient-based optimization, and hyperparameter tuning for large-scale image datasets (1 yr, 6 mos).</li> <li>Evaluate model inference performance and runtime behavior across heterogeneous hardware platforms, including high-performance computing (HPC) GPU environments, local NVIDIA GPU development systems, and VPU-based inference on production deployment machines, to ensure real-time execution