Applied Scientist , Amazon Customer Service

Amazon Amazon · Big Tech · Santa Clara, CA · Data Science

Applied Scientist II role focused on building AI-based automated customer service solutions using RAG, agentic AI, and post-training of LLMs. Responsibilities include designing and deploying RAG pipelines, conducting LLM post-training, curating datasets, implementing evaluation frameworks, developing AI agents, and collaborating with cross-functional teams. The role involves research and development with minimal guidance, aiming to translate research into production systems and contribute to the scientific community.

What you'd actually do

  1. Design, develop, and deploy information retrieval systems and RAG pipelines using embedding models, reranking algorithms, and generative models to improve customer service automation
  2. Conduct post-training of large language models using techniques such as Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Group Relative Policy Optimization (GRPO) to optimize model performance for customer service tasks
  3. Build and curate high-quality datasets for model training and evaluation, ensuring data quality and relevance for customer service applications
  4. Design and implement comprehensive evaluation frameworks, including data curation, metrics development, and methods such as LLM-as-a-judge to assess model performance
  5. Develop AI agents for automated customer service, understanding their advantages and common pitfalls, and implementing solutions that balance automation with customer satisfaction

Skills

Required

  • building models for business application experience
  • PhD, or Master's degree and 4+ years of CS, CE, ML or related field experience
  • Experience programming in Java, C++, Python or related language
  • Experience in any of the following areas: algorithms and data structures, parsing, numerical optimization, data mining, parallel and distributed computing, high-performance computing

Nice to have

  • Experience using Unix/Linux
  • Experience in professional software development

What the JD emphasized

  • build AI-based automated customer service solutions
  • retrieval-augmented generation (RAG)
  • agentic AI
  • post-training of large language models
  • Develop AI agents for automated customer service

Other signals

  • building AI-based automated customer service solutions
  • retrieval-augmented generation (RAG)
  • agentic AI
  • post-training of large language models
  • design, develop, and deploy information retrieval systems and RAG pipelines
  • Conduct post-training of large language models using techniques such as Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Group Relative Policy Optimization (GRPO)
  • Build and curate high-quality datasets for model training and evaluation
  • Design and implement comprehensive evaluation frameworks
  • Develop AI agents for automated customer service
  • Publish findings and contribute to the broader scientific community through papers, patents, and open-source contributions
  • Monitor and improve deployed models based on real-world performance metrics and customer feedback