Recommender Systems Unleashed

Join us for an engaging evening exploring the transformative impact of recommender systems across industries, from e-commerce and news to creative tools and beyond. Our speakers, leading experts from eBay, Taboola, and Lightricks, will share their experiences in designing innovative solutions that scale and adapt to unique challenges.
Guy Feigenblat will reveal how eBay personalizes the shopping journey for millions of global users. Yevgeny Tkach from Taboola will discuss crafting high-performance content recommendation systems for news platforms despite sparse data challenges. Asi Messica will delve into the creative applications of multimodal recommendations at Lightricks, where algorithms connect brands with creators and enhance user expression.
Whether you're a developer, data scientist, or simply fascinated by the technology driving personalization across industries, this meetup offers valuable insights and networking opportunities with fellow enthusiasts.

Timeline:
18:00- gathering

18:30-19:00- Recommending at scale - personalizing the eBay shopping journey - Guy Feignblat @ Ebay

19:00-19:30- Building Content Recommendation Systems for News Sites with Sparse Data at Scale- Yevgeny Tkach@Tabbola

19:30-20:00-When Recommender Systems Meet Creativity- Asi Messica -@Lightricks

Deatils
Recommending at scale - personalizing the eBay shopping journey - Guy Feignblat @ Ebay

Abstract:
eBay is a global online marketplace that connects millions of buyers and sellers, facilitating the exchange of goods across a vast range of categories. With over 2.1 billion items sold annually and 130 million active buyers across 190 marketplaces worldwide, eBay faces unique challenges in delivering personalized experiences at scale. In this talk, we will explore the advanced recommender systems developed by the Buyer Experience AI group at eBay, focusing on how we leverage Big Data to optimize the shopping experience. We will discuss the challenges of working with massive datasets, the methods used to extract user interests, and the techniques for personalizing recommendations to enhance the shopping journey. Additionally, we will delve into how we identify and support "good shopping missions," ensuring users find exactly what they need in an efficient and engaging manner.
Lecturer Bio:
Senior Manager at eBay R&D, where he leads the IL Buyer Experience AI group within the eBay Marketplace. In his current role, Guy oversees the applied research and engineering of recommender systems on eBay's Home Page, tackling complex challenges in AI, machine learning, and Big Data to improve the platform’s user experience.
Before joining eBay, Guy served as AI Director at Piiano. Earlier in his career, Guy was a Team Leader and Research Staff Member at IBM Research AI.He has authored numerous patents and published many papers in leading venues. Guy earned his Ph.D. in Computer Science from Bar-Ilan University.

Building Content Recommendation Systems for News Sites with Sparse Data at Scale - Yevgeny Tkach@Taabola

Abstract:
Delivering personalized content on large-scale platforms, such as news sites, presents unique challenges due to the vast number of users, items, and the sparsity of interaction data. This talk provides an in-depth look at the architecture of Taboola's recommendation systems, focusing on the two-stage process of candidation and reranking. We’ll explore how deep learning architectures are leveraged to tackle these challenges, achieving a balance between scalability and precision. Additionally, we’ll share insights and discuss the challenges faced when developing high-performance recommendation systems at scale.
Lecturer Bio
Machine learning team leader focusing on real time bidding algorithms to serve recommendations. Data scientist in the ad tech industry for the last 10 years**.**

When Recommender Systems Meet Creativity- Asi Messica @ Lightricks
Abstract:
Lightricks’ photo and video editing tools unlock endless possibilities for self-expression, while its creator services empower content creators to monetize their talents and collaborate with leading brands.
In this talk, we’ll provide an overview of the diverse recommender systems powering Lightricks’ products and take a deep dive into the algorithms and architecture of our multimodal recommender system. We will explore two distinct use cases, each with its own unique constraints and challenges: matching brands with creators who align with their values and campaign goals, and recommending the most suitable features for a specific image a user has uploaded.

Lecturer Bio
Dr. Asi Messica is VP Data Science at Lightricks and a lecturer at Reichman University. Before joining Lightricks, she held various data science, product, and development management positions at Fiverr, SAP, RSA, and more.
She pursued her Ph.D. in the area of Recommender Systems at Ben-Gurion University. Her professional interests include machine learning, personalization, information retrieval, NLP and reinforcement learning.