AI/ML Engineer
Mid-level computer vision engineer on a three-person pod building custom YOLO object detection and classification models that process live and recorded full-motion video for U.S. Customs and Border Protection. Day to day: curate and annotate datasets, train and tune PyTorch/TensorFlow models, and build Python inference pipelines, on-site 2-3 days a week in Ashburn, VA.
- Compensation
- Not disclosed
- Location
- Ashburn, United States
- Work setup
- Not specified
- Seniority
- Middle
- Employment
- Not specified
- Experience
- 2+ years
- Category
- Ml / AI
- Posted
Skills
Must have
- B.S. in Computer Science, Data Science, Software Engineering, or related discipline
- 5+ years software engineering or data science experience, including 2+ years building and training computer vision models
- Python programming with core ML libraries (NumPy, Pandas, Scikit-Learn)
- Hands-on experience training neural networks with PyTorch or TensorFlow
- Experience implementing YOLO (e.g., YOLOv8/v9) for object detection, segmentation, or classification
- Image and video processing experience with OpenCV or FFmpeg
- Data curation and formatting, with experience using annotation platforms such as CVAT or LabelImg
- Git version control and collaboration on shared codebases
- Strong analytical skills with focus on benchmarking model performance (mAP, FPS, inference latency)
- Willing to work on site 2-3 days per week in Ashburn, VA
- U.S. Citizenship and ability to pass a CBP background investigation
Nice to have
- Containerized development environments (Docker)
- GPU execution environments and model export formats (ONNX, TensorRT)
- Multi-object tracking libraries
- Agile tools such as Jira and Confluence
Full description
CDIT is
seeking a Mid-Level AI/ML Computer Vision Engineer to support the BEAGLE Agile
Solution Factory (ASF) team serving U.S. Customs and Border Protection (CBP) in
Northern Virginia.
You will be
part of a focused three-person computer vision pod building custom object
detection and classification models with the YOLO framework. The team processes
live and recorded Full-Motion Video (FMV) feeds to automatically identify
roads, vehicles, people, and terrain features. This is a hands-on role: you
will bring raw production video into a secure training environment, curate
datasets, train and tune models, and deploy high-performance inference
pipelines.
**Key
Responsibilities**
- Model Training & Experimentation: Train, fine-tune, and evaluate custom YOLO object detection models on
proprietary datasets; experiment with hyperparameters, augmentations, and model
sizes to balance speed and accuracy.
- Data Engineering & Ingestion: Process, clean, and format raw video recorded in production
environments into high-quality training datasets.
- Dataset Annotation & Curation: Manage the annotation lifecycle, including labeling images and video
frames, verifying annotation quality, and generating synthetic data when
needed.
- Inference Pipeline Development: Write Python scripts that parse video frames, feed them into inference
engines, and export structured prediction data (JSON/Protobuf) for detected
objects.
- Model Evaluation: Produce performance reports, confusion matrices, and precision-recall
curves to find model blind spots and drive continuous improvement.
- Agile Delivery: Partner with the Senior AI/ML Lead, video engineers, and developers in
an Agile/Scrum environment to deliver production-ready code in short sprints.
**Required
Qualifications**
- B.S. in Computer Science, Data Science,
Software Engineering, or a related discipline.
- 5+ years of software engineering or data
science experience, including 2+ years hands-on building and training computer
vision models.