Applied Scientist Ii, Aft Ai, Amazon Aft AI

Amazon Amazon · Big Tech · DE, Belgium +1 · Applied Science

Applied Scientist II role focused on developing and deploying agentic AI solutions and multi-modal deep learning models for Amazon's Fulfillment Network. The role involves working with large-scale, real-world datasets (imagery, natural language, structured data) to solve complex problems like warehouse operations and visual defect detection, pushing the state-of-the-art in optimizing fulfillment systems.

What you'd actually do

  1. build agentic AI solutions and multi-modal deep learning models that understand how products and packages flowing through Amazon’s fulfillment network.
  2. build models that solve challenging problems like understanding warehouse operations systems, or visual defect detection on Amazon's entire retail catalog (billions of different items, thousands of new items every day).
  3. work with a diverse set of very large multi-modal real-world datasets, including imagery, natural language and structured data.
  4. face a high level of research ambiguity and problems that require creative, ambitious, and inventive solutions.
  5. adapt state-of-the-art agentic AI, deep learning, language understanding and computer vision techniques to develop solutions for business problems in the Amazon Fulfillment Network.

Skills

Required

  • PhD, or a Master's degree and experience in solving business problems through machine learning, data mining and statistical algorithms
  • Experience in building models for business application
  • Experience in patents or publications at top-tier peer-reviewed conferences or journals
  • Strong programming proficiency in Python with production-quality code standards; deep technical expertise with PyTorch and proficiency with the modern ML stack (Pandas, NumPy, scikit-learn, Hugging Face Transformers)
  • Demonstrated ability to design and execute end-to-end ML projects from research through production deployment, with experience in model monitoring and iterative improvement
  • Strong expertise in modern deep learning architectures including transformers and diffusion models, with hands-on experience in training optimization techniques (distributed training, mixed precision, gradient accumulation) and model compression methods (quantization, pruning, distillation)
  • Experience fine-tuning large language models (GPT, LLaMA, Claude) and vision-language models (CLIP, LLaVA, Qwen)
  • Proven experience developing agentic AI systems using state-of-the-art frameworks (LangChain, Strands, etc.) with ability to design tool-augmented reasoning systems, RAG systems, and advanced prompt engineering techniques (chain-of-thought, few-shot)
  • Strong knowledge and hands-on experience across multiple ML domains including computer vision (object detection, segmentation, classification), natural language processing (text generation, information extraction), and multimodal learning
  • Understanding of ML systems design including model serving infrastructure, A/B testing frameworks, and MLOps best practices

Nice to have

  • Experience in professional software development
  • Experience with explainable machine learning and artificial intelligence methodologies and tools
  • Experience working with large language models (GPT, LLaMA, Claude) and vision-language models (CLIP, LLaVA, Qwen) in production settings
  • Experience collaborating on cross-functional ML initiatives with demonstrated impact on product metrics
  • Multiple publications in top-tier venues, including co-authored papers or contributions to ML research communities
  • Experience with generative AI techniques including diffusion models for image/video synthesis, autoregressive models for multimodal generation, and controllable generation systems
  • Experience with specialized ML domains such as few-shot learning, meta-learning, or domain adaptation; ability to build models that handle distribution shifts or long-tail scenarios

What the JD emphasized

  • production-quality code standards
  • end-to-end ML projects from research through production deployment
  • model monitoring and iterative improvement
  • training optimization techniques
  • model compression methods
  • fine-tuning large language models
  • vision-language models
  • agentic AI systems
  • tool-augmented reasoning systems
  • RAG systems
  • advanced prompt engineering techniques
  • computer vision
  • natural language processing
  • multimodal learning
  • ML systems design
  • model serving infrastructure
  • A/B testing frameworks
  • MLOps best practices

Other signals

  • applying state-of-the-art AI on real-world problems at truly vast scale
  • build and deploy models that make smarter decisions on a wide array of multi-modal signals
  • pushing beyond the state of the art in optimizing one of the most complex systems in the world: Amazon's Fulfillment Network
  • build agentic AI solutions and multi-modal deep learning models
  • visual defect detection on Amazon's entire retail catalog
  • work with a diverse set of very large multi-modal real-world datasets
  • high level of research ambiguity and problems that require creative, ambitious, and inventive solutions
  • adapt state-of-the-art agentic AI, deep learning, language understanding and computer vision techniques
  • develop solutions for business problems in the Amazon Fulfillment Network
  • build models that solve challenging problems like understanding warehouse operations systems