Distinguished Scientist, Agentic - Traveler

Why Join Us?

To shape the future of travel, people must come first. Guided by our Values and Leadership Agreements, we foster an open culture where everyone belongs, differences are celebrated, and when one of us wins, we all win.

We provide a full benefits package, including exciting travel perks, generous time‑off, parental leave, a flexible work model (with some pretty cool offices), and career development resources, all to fuel our employees' passion for travel and ensure a rewarding career journey. We’re building a more open world. Join us.

Expedia Product & Technology builds innovative products, services, and tools to deliver high‑quality experiences for travelers, partners, and our employees. A singular technology platform powered by data and machine learning provides secure, differentiated, and personalized experiences for the traveler and our partners that drive loyalty and customer satisfaction.

The AI & Data Science team is breaking new ground to tackle some of the most complex customer experience problems in the travel domain. We lead the science, architecture, and experimentation behind traveler‑facing agentic experiences, spanning Vrbo hosts, hotels, and other supply partners. This team will shape how autonomous AI agents help millions of travelers plan, book, and manage their trips, safely and at scale—for example, turning requests like "Plan a 10‑day family trip to Japan with safe hotels and direct flights" or "Find me a villa with a pool under $300/night" into secure, personalized, and actionable itineraries across our B2C platforms.

• Shape the long‑term vision and scientific agenda for agentic applications serving travelers, identifying “big‑hairy” opportunities where autonomous agents can transform the end‑to‑end trip lifecycle.
• Serve as a strategic thought partner to senior product, engineering, and platform leaders, defining the role of agentic interfaces vs. traditional search/GUI tooling and ensuring we deliver real business value while preserving trust and a delightful user experience.
• Design and evolve agentic experiences and architectures so that natural language from travelers is mapped to structured intents and tool calls that safely interact with our shopping and booking platforms.
• Lead applied research in areas such as complex itinerary planning, ambiguity resolution, and evaluation methodologies.
• Be hands‑on: prototype agents and components, design and run experiments, build and evaluate models, and work with engineering teams on end‑to‑end systems.
• Lead end‑to‑end experimentation, from ideation and offline evaluation to A/B tests in production.

• Bachelor of Science degree in Computer Science, Machine Learning, or a related field, with 15+ years of experience.
• Deep expertise in LLM training, tuning, distillation, multi‑agent frameworks, and orchestration of tool‑using agents, along with architecting ML platforms and systems to support large‑scale experimentation and deployment of agents.
• Familiarity with agentic evaluation frameworks (e.g., Lang Smith, Deep Eval, Galileo), agentic protocols (e.g., MCP, A2A, AG‑UI, ACP), and agentic frameworks (e.g., Lang Graph, Llama Index, Auto Gen), or comparable internal/industry tools.
• Demonstrated ability to deliver large, complex GenAI projects from concept to production.
• Broad and deep understanding of machine learning theory and practice, including recommender systems, personalization, and contextual engineering patterns (retrieval, memory, session/state management) and how they interface with agentic applications.
• Proficiency in one major ML programming language (e.g., Python) and familiarity with data pipelines and model deployment tooling and practices.

• PhD in Computer Science, Machine Learning, or a related field with 12+ years’ experience.
• An entrepreneurial mindset, with a history of converting proofs‑of‑concept into revenue‑generating solutions and tangible business outcomes.
• Broad and deep understanding of machine learning theory and practice across multiple sub‑domains, comfortable operating as a “jack of all trades” across modeling, data,…