Principal Data Scientist, Reinforcement Learning

Expedia Group
Expedia Group

Job Overview

Expedia Group is seeking a specialist in real world Reinforcement Learning for a high-level role putting contextual bandits and wider RL techniques to productive use across all of our brands and platforms.

This is a high-value role for Expedia Group focused on building a core system for the creation of adaptive user experiences intended to drive significant gains in engagement and conversion via personalizing web pages, mobile apps, emails and other critical customer touchpoints.

What you’ll do:

This is a technical role requiring outstanding skills and experience both in terms of the theory of reinforcement learning and the applied side of building and scaling an RL system in production. Competitive candidates are expected to contribute at a very high-level as a data scientist, software engineer and ML researcher. Additionally, candidates should have excellent communication skills, strong product sense, an ability to prioritize and be very self-motivated. This role is expected to serve as an internal leader of the Expedia RL community and of Expedia’s efforts to publish external papers.

Machine Learning Engineering:

  • Reinforcement Learning

  • Including contextual and context-free bandits

  • Off-policy evaluation

  • Ability to navigate a complex platform and ensure that technical decisions are made with an eye toward simplicity but also long-term growth and scalability

  • Deep experience with programming languages such as Python, Java, Scala

  • Experience with common data tools: Spark, SQL, Kafka, Flink

  • Popular ML tools: Vowpal Wabbit, PyTorch, Scikit Learn, Tensorflow RL

  • Experience defining and testing complex objective functions in partnership with business partners

  • Experience building and scaling consumer facing ML systems at low latency and high throughput

  • Understanding of modern software design and testing: Devops, CI/CD, Unit testing

  • A history of careful but ambitious A/B testing and foundations in experimental design

  • Statistics:

  • You will have a solid theoretical foundation and apply advanced statistical methods to a broad range of problems.

  • You have experience in advanced experimental design (e.g. adaptive designs).

  • You will read relevant publications in the field and implement the methods described to business problems.

  • You will act as a technical mentor for junior profiles.

  • Model Design:

  • You will build end-to-end design of one step models based on detailed understanding of business requirements, including general approach, choice of algorithm and data sources.

  • Your model output reflects a deep understanding of business requirements adhering to standard methodologies, Academic research and experience.

  • You have expert knowledge of complex model design incorporating the latest academic and published research including neural network, transformers and reinforcement learning domains.

  • You will display advanced domain (i.e. travel, online retail) knowledge, business acumen (understanding the underlying business objectives), and critical reasoning skills.

  • You will demonstrate a critical understanding of business processes and challenges to recommend and develop solutions to meet unique business needs and achieve objectives.

  • Communication and presentation:

  • You will be a persuasive storyteller using influencing techniques that build trust with team and partners and establish long-term relationships and connect the vision, strategy and priorities across brands.

  • You will contribute ideas and participates in industry events and forums and are recognized as an inspiring leader outside of the company.

  • You will drive a culture of constructive critique, challenging others to bring fresh and creative perspectives to differentiating functional technical communications.

  • You will creatively develop tools, visual aids and other original content to support the understanding/education of sophisticated and technical information (e.g. legal, operational risk, and regulatory data).

  • You will make recommendations for business process improvement, sponsoring and staging workshops/ knowledge sessions to improve the organization’s functional knowledge and acuity, and leveraging insights and subject matter experts to overcome challenges.

  • Who you are:

  • MS/MSc or PhD in computer science, electrical engineering, statistics or similarly quantitative field

  • Proven experience creating and growing large-scale production learning systems

  • Publications in relevant venues: KDD, COLT, NeurIPS, ICLR, RecSys, ICML, SIGIR

  • Bonus points for contributions to relevant open source projects

  • Some background solving hard problems in industrial ML: computational advertising, large-scale recommender systems, fraud and anomaly detection in non-stationary environments, click-through optimization in eCommerce, etc

  • View More
    Job Detail
    Shortlist Never pay anyone for job application test or interview.