Applied Scientist, Conversational Assistant Modeling and Learning

Amazon; San Francisco (CA); On-site; USD 142,800 - 193,200; Full time

Benefits offered by this job

Comprehensive health insurance

401(k) matching

Paid time off

Job summary

Amazon is seeking an Applied Scientist in San Francisco to work on next-gen conversational AI with Alexa+. This role demands deep expertise in LLMs and offers the chance to lead groundbreaking technical initiatives that impact millions.

The successful candidate will require extensive knowledge in programming, distributed training, and machine learning, along with qualifications such as a Master’s degree in relevant fields. Join us in shaping the future of conversational technology!

Qualifications

• 3+ years of programming experience.
• 1+ year of experience with distributed training frameworks.
• Solid ML background, familiar with NLU and LLM training.

Responsibilities

• Improve efficiency of LLM training and evaluation pipelines.
• Design and implement fine-tuning and alignment approaches.
• Architect agentic systems for multi-step reasoning.

Skills

Programming in Java, C++, Python

Distributed training frameworks (VeRL, Megatron, etc.)

Machine learning tools (PyTorch, TensorFlow)

Natural Language Understanding (NLU)

Natural Language Generation (NLG)

Education

Master's degree or above in Computer Science, Machine Learning, or Engineering

Tools

PyTorch

TensorFlow

Job description

Applied Scientist, Conversational Assistant Modeling and Learning

Job ID: 10449183 | Amazon.com Services LLC

Alexa AI is looking for an Applied Scientist to build Alexa+, Amazon's LLM-powered conversational assistant. You will work on key initiatives spanning large language model fine-tuning, alignment, agentic reasoning, and evaluation — directly shaping the experience for hundreds of millions of customers worldwide.

A successful candidate will be a self-starter comfortable with ambiguity, strong attention to detail, and the ability to work in a fast‑paced, ever‑changing environment. As an Applied Scientist, you will own the design and development of end‑to‑end systems. You’ll have the opportunity to create technical roadmaps, and drive production‑level projects that will support Amazon Science. You will work closely with other scientists and engineers to develop solutions and deploy them into production. The ideal scientist must have the ability to work with diverse groups of people and cross‑functional teams to solve complex business problems.

Key Job Responsibilities

• Improve the efficiency of LLM, VLM, and agent training and evaluation pipelines, including distributed training, inference serving, data loading, checkpointing, memory usage, and GPU utilization.
• Design, implement, and evaluate novel approaches to LLM fine‑tuning, alignment (RLHF, DPO), and distillation for production deployment.
• Architect agentic systems — multi‑step reasoning, tool use, planning, and orchestration.
• Develop evaluation frameworks and methodologies that go beyond standard benchmarks to capture real‑world conversational quality.
• Translate research advances into customer‑facing products, working closely with engineering, product, and cross‑functional science teams.
• Publish results at top‑tier venues and represent Amazon in the broader research community.

About the Team

Alexa AI is building the science and technology behind Alexa+, Amazon's next‑generation conversational assistant. Our team works at the intersection of large language models, reinforcement learning, agentic architectures, and multilingual/multimodal understanding. We operate at massive scale — our models serve customers across dozens of languages and device types. If you want to push the frontier of conversational AI and see your work used by people every day, come join us.

Basic Qualifications

• Master's degree or above in computer science, machine learning, engineering, or related fields.
• 3+ years of programming in Java, C++, Python or related language experience.
• 1+ years experience with distributed training frameworks such as VeRL, Megatron, FSDP, DeepSpeed, Ray, or similar systems, and inference engines such as vLLM, TensorRT‑LLM, Triton, SGLang, TGI.
• 3+ years’ experience with modeling languages and tools like PyTorch / TensorFlow, R, scikit‑learn, numpy, scipy, etc.
• Solid ML background and familiar with NLU, NLG, and LLM training and evaluation.

Preferred Qualifications

• PhD in computer science, machine learning, engineering, or related fields.
• 3+ years experience with distributed training frameworks such as VeRL, Megatron, FSDP, DeepSpeed, Ray, or similar systems, and inference engines such as vLLM, TensorRT‑LLM, Triton, SGLang, TGI.
• Publications at peer‑reviewed NLP/ML conferences (e.g. ACL, EMNLP, NAACL, NeurIPS, ICLR, ICML, AAAI, etc.).
• Scientific thinking and the ability to invent, a track record of thought leadership and contributions that have advanced the field.

Benefits & EEO Statement

Amazon is an equal opportunity employer and does not discriminate on the basis of protected veteran status, disability, or other legally protected status. If you have a disability and need a workplace accommodation or adjustment during the application and hiring process, please consult the company’s accommodations page. Amazon offers comprehensive benefits including health insurance (medical, dental, vision, prescription, Basic Life & AD&D insurance, and options for Supplemental life plans, EAP, Mental Health Support, Medical Advice Line, Flexible Spending Accounts, Adoption and Surrogacy Reimbursement coverage), 401(k) matching, paid time off, and parental leave.

Compensation

USA, WA, Bellevue − $142,800.00 − $193,200.00 USD annually.