Applied Scientist: ML & Vision for Production Systems

Amazon; Reading (MA); On-site; USD 136,000 - 184,000; Full time

Benefits offered by this job

Health insurance

401(k) matching

Paid time off

Parental leave

Job summary

Amazon is seeking an Applied Scientist on the Science SW team to serve as a versatile generalist, framing ambiguous business problems as tractable scientific problems and delivering ML systems from prototype to production across perception, optimization, and orchestration.

You will collaborate with researchers and engineers to prototype edge-deployable models, validate sensing hardware, and integrate solutions into scalable deployment architectures that improve fulfillment performance.

Qualifications

• PhD or master’s/PhD-equivalent in a quantitative field with strong ML/data science background.
• Experience with Python/Java/C++ in ML production environments.
• Strong foundations in statistics, optimization, and machine learning across domains.
• Experience with PyTorch or TensorFlow and scientific Python stack.

Responsibilities

• Own research and development of ML solutions across ML, CV, optimization, and physics-informed modeling in production.
• Ramp quickly on new domains by framing problems as tractable scientific questions and prototyping end-to-end.
• Prototype and evaluate edge-deployable models and sensing hardware under real-world constraints.
• Collaborate across science and engineering teams to integrate solutions into deployment architectures.

Skills

Python

Java

C++

Applied mathematics

ML theory

Education

Advanced degree in engineering/CS/quantitative field

Tools

PyTorch

TensorFlow

NumPy

SciPy

scikit-learn

pandas

Job description

Amazon is seeking an Applied Scientist on the Science SW team to serve as a versatile generalist, framing ambiguous business problems as tractable scientific problems and delivering ML systems from prototype to production across perception, optimization, and orchestration.

You will collaborate with researchers and engineers to prototype edge-deployable models, validate sensing hardware, and integrate solutions into scalable deployment architectures that improve fulfillment performance.