Engineering Blog

Our goal: Help developers create better code, more efficiently, while maintaining accessibility for all.

By leveraging Apache Spark, eBay Graph Database can export a graph with billions of vertices and edges.

We developed a ranking model to generate personalized recommendations that optimally ranks hundreds of candidate recommendations by considering a user’s shopping experience as well as platform performance objectives.

By integrating information from different modalities of eBay listings such as titles and images, we greatly improved the buyer experience and relevance of the recommended items on eBay’s listing pages.

Chief AI Officer Nitzan Mekel-Bobrov and Vice President of Seller Experience Xiaodi Zhang shared how AI will make life easier for sellers, buyers and employees.

Mazen shared his perspective on emerging technology trends and how eBay is embracing the disruption, bringing new customer experiences powered by AI to the marketplace.

Here at eBay, we’ve crafted a brand new approach to automate platform evolution for all applications — one that provides a repeatable and reusable infrastructure to streamline evolution.

The Similarity Engine's use cases include item-to-item similarity for text and image modality and user-to-item personalized recommendations based on a user’s historical behavior data.

eBay's core page — the View Item page — gets 250 million views per day. Here's how we took on the task of modernizing it.

A new feature generates customer delight by using modern computer vision techniques to drive new search paradigms through visual discovery.

It’s been effective at speeding up productivity, pushed to eBay, and contributed back to the open source community.

A new risk data hub for decisioning can significantly boost variable time-to-market, with high performance under billions of query traffic.

Contract testing has grown in popularity in recent years with the widespread adoption of microservice architectures. In this article, we will share our experiences with contract testing at eBay.

By leveraging deep learning techniques to compare the titles of product listings, we greatly improved the relevance of our recommended items on eBay’s View Item page.

eBay’s notification platform team built a fault-tolerant, resilient system by injecting faults in the application level.

eBay makes a crucial pivot to OpenTelemetry to better align with industry standards for Observability.

How the powerful meta-analysis method called "weighted z-test" can help eBay become more efficient.

Sellers in the U.S., U.K., Germany, Australia, France, Italy and Spain now have access to Buyer Groups, a new tool which allows them to segment their buyers to drive more repeat business.

Determining which promoted auction items to display in a merchandising placement is a multi-sided customer challenge that presents opportunities to both surface amazing auction inventory to buyers and help sellers boost visibility on their auction listings.

Billions of queries require new, smarter alerting features.

A case study demonstrates how eBay's Notification Engineering team optimizes a streaming system in a microservice architecture to support high-throughput broadcast notifications.

Buyers reveal a whole range of behaviors and interests when they browse our pages, so we decided to incorporate these additional purchase intent signals into our machine learning model to improve the relevance of our recommended items.

The new open feature flagging standard enables companies to deliver cloud-native applications more effectively.

For our annual ML competition, we challenged university students to predict how many days a carrier takes to deliver packages.

Under the new machine learning model, buyers are recommended items that are more aligned to their shopping interests on eBay.

A step-by-step guide on how to build a state-of-the-art recommender system in an industrial setting.

A look at how our eBay technologists created an automatic testing solution for batch applications.

At eBay, we’ve focused on improving our staging environment to increase our software delivery speed, which ultimately allows us to launch features faster to our customers. Over the last several months, we’ve raised the quality of our staging data by establishing a dedicated and scalable data migration system.

A new lightweight distributed architecture is proposed and applied on eBay’s release system to support thousands of libraries release work in high efficiency.