Edge ML Engineer — Computer Vision & Object Tracking

Motion Recruitment; Arlington (VA); Hybrid; USD 150,000 - 210,000; Full time

Job summary

Motion Recruitment is seeking a senior applied machine learning engineer to join a high-performing CV team in Arlington. You will develop and deploy object detection and multi-object tracking models across real-world applications, owning the full ML lifecycle from data curation to production deployment.

You will optimize models for edge hardware and work with TensorRT/ONNX to meet latency and compute constraints.

Qualifications

• 5+ years of professional experience in applied machine learning or computer vision.
• Bachelor’s degree in Computer Science, Electrical Engineering, or related field; Master’s/PhD valued.
• Proven experience deploying object detection models in production environments.
• Strong Python programming with PyTorch or similar framework.
• Deep understanding of CV datasets, annotation quality, training behavior, and evaluation metrics.
• Experience end-to-end ML lifecycle: data, annotation, training, evaluation, optimization, deployment.

Responsibilities

• Develop, train, and fine-tune object detection models (YOLO, DETR, Faster R-CNN).
• Build and improve multi-object tracking pipelines (SORT, DeepSORT, ByteTrack).
• Evaluate model performance, analyze failures, and propose improvements.
• Own data and model pipeline: data curation, annotation coordination, augmentation, training, validation.
• Optimize models for edge/embedded hardware balancing speed, accuracy, and compute.
• Collaborate with software/systems teams to integrate CV models into production.

Skills

Python programming

PyTorch

Object detection / CV

Multi-object tracking

Edge deployment

Model evaluation

MLOps basics

Education

Bachelor's degree in Computer Science/EE

Tools

TensorRT

ONNX

CUDA

CVAT / Label Studio

Job description

Join a high-performing and growing applied machine learning team building advanced computer vision capabilities for real-world, mission-critical applications. In this role, you’ll focus on developing and deploying object detection and multi-object tracking models, working with complex datasets and bringing models from initial experimentation through production deployment.

This is a highly hands-on position where you’ll own the full ML lifecycle, including data curation and annotation, model training and evaluation, performance optimization, and deployment to edge and embedded platforms. You’ll work closely with software and systems engineers to ensure models perform reliably in challenging real-world environments where accuracy, latency, and compute limitations all matter.

What You’ll Do

• Develop, train, and fine-tune object detection models using architectures such as YOLO, DETR, Faster R-CNN, or similar approaches.

• Build and improve multi-object tracking pipelines using frameworks and techniques such as SORT, DeepSORT, ByteTrack, or comparable methods.

• Evaluate model performance by analyzing metrics, identifying failure cases, and determining where improvements can be made across both the model and underlying data.

• Own the end-to-end data and model pipeline, including reviewing raw data, coordinating annotation efforts, curating datasets, implementing augmentation strategies, training models, and validating results.

• Optimize deep learning models for edge and resource-constrained hardware, balancing inference speed, accuracy, memory, and compute limitations.

• Utilize technologies such as TensorRT, ONNX, quantization, and pruning to improve model performance for production environments.

• Collaborate with software and systems engineering teams to integrate computer vision models into broader production systems.

• Work with imagery across different sensor modalities, including electro-optical, infrared, thermal, and other imaging sources.

• Continuously improve deployed computer vision systems based on real-world performance and new datasets.

What We’re Looking For

• 5+ years of professional experience in applied machine learning, computer vision, or a related technical discipline.

• Bachelor’s degree in Computer Science, Electrical Engineering, or a similar technical field. A Master’s degree or PhD is highly valued.

• Proven experience developing and deploying object detection models in production environments, rather than exclusively academic or research-based work.

• Strong programming skills in Python with hands-on experience using PyTorch or another modern deep learning framework.

• Strong understanding of computer vision datasets, annotation quality, training behavior, and evaluation metrics.

• Ability to diagnose why a model is underperforming and determine whether improvements need to come from the model architecture, training process, dataset, annotations, or other factors.

• Experience managing the complete ML development lifecycle, including data preparation, annotation, training, experimentation, evaluation, optimization, and deployment.

• Hands-on experience optimizing ML models for edge, embedded, or compute-constrained environments.

• Familiarity with technologies such as TensorRT, ONNX, model quantization, and other inference optimization techniques.

• Practical understanding of multi-object tracking with experience implementing or working with tracking algorithms.

• Ability to understand current computer vision research and translate relevant techniques into practical improvements for production systems.

• U.S. citizenship is required (must be able to obtain a clearance).

Nice to Have

• Strong development experience with C++ or Rust, particularly for production ML systems, inference, or performance optimization.

• Experience working with infrared, thermal, or other specialized sensor data.

• Familiarity with MLOps and experiment management tools, such as MLflow, Weights & Biases, dataset versioning platforms, or model registries.

• Experience managing computer vision annotation workflows using tools such as CVAT, Label Studio, or comparable platforms.

• Broader computer vision experience involving areas such as segmentation, pose estimation, or activity recognition.

• Experience leveraging simulation, emulation, or synthetic data for model development, training, or evaluation.

• Experience deploying machine learning models onto GPU-accelerated embedded hardware, such as NVIDIA Jetson or similar platforms.

• Previous experience working within defense, intelligence, aerospace, autonomous systems, robotics, or other mission-critical environments.