Data Science Intern

Airbnb

5h•Remote

About The Position

Airbnb was founded in 2007 and has grown to over 5 million hosts and 2 billion guest arrivals globally, connecting guests with communities through unique stays and experiences. The company is seeking a (PhD) Data Science Intern to join the Core Data Science team for its 2026 Summer Intern Program, running from June 1, 2026, to August 21, 2026. The Data Science organization aims to be a business multiplier by addressing complex technical questions through data research, modeling, and empirical methods. Core Data Science specifically focuses on innovating and applying deep scientific approaches to critical problems on the Airbnb platform, accelerating scalable scientific innovation by introducing new methods and technologies across the company. During the internship, the intern will be integrated into the Trip Intelligence working group, which seeks to enhance understanding of real-world trip experiences. The primary goal of this role is to improve the quality of the GSAT model, which infers Guest satisfaction using AI applications on various data sources. This internship provides a unique opportunity to work on a cutting-edge problem, utilizing GenAI methods for exploratory research that directly impacts the core Airbnb experience and offers valuable field experience.

Requirements

Full-time Doctorate students enrolled at a nationally-accredited university in the United States graduating between December 2026 - June 2027
Studying computer science, computer engineering, data science, or an equivalent technical field
Knowledge of cutting edge GenAI techniques in the image domain
Ability to read and summarize papers from the literature
Solid Python coding skills
Familiarity with at least one DNN framework (PyTorch or Tensorflow)
Demonstrated interest in high growth, technology, and/or hospitality companies
Work authorization for employment in the United States is required (CPT/OPT with 2-year STEM extension is accepted)

Responsibilities

Conducting literature reviews to identify ways in which active learning can be integrated into our current model
Investigating innovative approaches to improving labeling for training and evaluation datasets
Implementing an active learning pipeline for our existing model and analyzing the performance
Investigating other approaches to improving label quality (e.g. LLM labeling agents)