Drimva

Expedia

Senior Machine Learning Scientist, Agentic AI & GenAI Systems (Growth Marketing)

$173K–$277K

Washington - Seattle Campus · Senior · Full-time · Posted Oct 30, 2026

Experience
10+ years in software engineering, ML, and AI systems, with production GenAI deployments.

About the job

Build and lead production-scale agentic and generative AI systems for personalized travel discovery, marketing campaigns, and user-facing products.

What you'll do

  • Architect, implement, and deploy end-to-end GenAI, RAG, and multi-agent systems, including interfaces, memory, evaluation, observability, and safety features.
  • Train, adapt, fine-tune, and distill LLMs; build multimodal pipelines and large-scale behavioral embedding systems for personalization and marketing.
  • Lead technical standards and design reviews, mentor team members, and collaborate with product and engineering teams to deliver user-facing experiences.

What they ask for

  • Deep expertise in LLM training, adaptation, distillation, RLHF/DPO, and RAG systems.
  • Strong foundation in NLP and experience with multimodal AI systems, including vision-language models.
  • Experience building and operating multi-agent AI platforms with observability and safety frameworks.
  • Background in distributed GPU training and inference, cloud infrastructure (AWS/Azure), container orchestration, and ML tooling.

Nice to have

  • PhD in Computer Science, Machine Learning, or a related field.
  • Industry presence through publications, patents, talks, or open-source contributions in LLMs, RAG, or agentic systems.

Pay

$173,000.00 to $242,500.00 USD total cash; up to $277,000.00 USD total cash (top of range)

Tools

LangChain, LangSmith, AutoGen, OpenAI/Claude Agents SDK, LangGraph, Claude, AWS, Azure

AI focus

Production GenAI and agentic AI for marketing and travel discovery, including LLM training and adaptation, RAG, multi-agent systems, multimodal AI, behavioral embeddings, personalization, evaluation, safety, and monitoring.

Model-written summary · the original posting is authoritative

First seen Oct 31, 2026 · Last verified Nov 5, 2026 (UTC)