Aerial Vision ML Engineer: Detection & Classification

Raad Labs; United States; On-site; USD 180,000 - 230,000; Full time

Benefits offered by this job

Competitive pay & equity

Health, dental & vision

Remote-first

Gear & home office budget

Generous time off

Learning & travel

Job summary

Raad Labs is seeking a Machine Learning Engineer to advance aerial image classification. You will train and ship detection/segmentation models, curate aerial datasets, and deploy robust pipelines from research to production.

You should have 4+ years of CV model experience, deep PyTorch expertise, and strong Python engineering habits. The role offers remote work within the USA, competitive compensation, equity, and an emphasis on scalable, edge-friendly inference.

Qualifications

• 4+ years building and shipping computer vision models to production.
• Proficient in PyTorch and modern detection literature and toolchains.
• Strong Python engineering habits; production-grade training code.

Responsibilities

• Train, evaluate, and ship detection and classification models (YOLO-family, ViT-based, segmentation) for aerial inspection.
• Own dataset curation and augmentation for aerial-specific challenges.
• Build evaluation harnesses with per-class, per-region, per-sensor breakdowns.
• Work in Python across the stack: PyTorch, ONNX/TensorRT export, tooling for annotation, training, deployment.
• Partner with edge team to quantize and prune models for on-device inference.

Skills

4+ years CV models

PyTorch

Python development

Model optimization

Edge deployment

Tools

PyTorch

ONNX

TensorRT

ONNX Runtime

Quantization

Job description

Raad Labs is seeking a Machine Learning Engineer to advance aerial image classification. You will train and ship detection/segmentation models, curate aerial datasets, and deploy robust pipelines from research to production.

You should have 4+ years of CV model experience, deep PyTorch expertise, and strong Python engineering habits. The role offers remote work within the USA, competitive compensation, equity, and an emphasis on scalable, edge-friendly inference.