Recommendation Large Model Researcher, Global E-commerce, Soaring Star Talent Program

Role Overview

This position sits within the international e-commerce recommendation team, which powers personalized discovery across major surfaces such as the mall homepage, purchase funnels, product detail pages, and store/showcase experiences. The team supports a massive daily user base with tailored suggestions for products, live streams, and short-form video content, while focusing on solving advanced recommendation-system problems through algorithmic improvements.

Project Focus

The assignment is centered on investigating new large-model approaches for recommendation, with the goal of moving beyond established recommendation architectures and infrastructure patterns. The objective is to build solutions that outperform current baseline systems and can be extended across multiple business products, including short video, LIVE, e-commerce, and news feeds.

Key Technical Challenges

Building large recommendation models presents several practical obstacles, including very high expectations for engineering efficiency, the need to preserve strong personalization, and the difficulty of representing content effectively across formats such as short video and live-stream media. The work will explore how to overcome these issues through research and system design.

The main research directions include parameter scaling for recommendation models, better learning of user and content representations, multimodal understanding, modeling very long behavior sequences, and generative recommendation methods.

Research Areas

• Representation learning driven by content understanding and user behavior
• Scaling model parameters and computation for recommendation systems
• Modeling ultra-long user sequences
• Generative approaches for recommendation

Qualification Requirements

• A doctoral degree is required, with preference for candidates in computer science, mathematics, or closely related disciplines.
• A strong base in machine learning and programming is expected, along with research experience in machine learning, NLP, computer vision, or related areas.
• Applicants should have solid knowledge of core algorithms and data structures.
• Experience contributing to or leading major projects in search, advertising, recommendation, or large-model work is an advantage.
• Preference will be given to candidates with publications in leading international conferences such as KDD, SIGIR, RecSys, ACL, and NeurIPS.
• Strong analytical thinking, problem-solving ability, curiosity, and enthusiasm for tackling difficult technical challenges are important.

About the Company

Founded in 2012, ByteDance builds products designed to inspire creativity and enrich life. Its portfolio includes global products such as TikTok, Lemon8, CapCut, and Pico, along with China-market products such as Toutiao, Douyin, and Xigua.

Why This Company

The organization emphasizes creativity, impact, and continuous iteration. Teams are global and diverse, and the culture values curiosity, humility, and an “Always Day 1” mindset. Employees are encouraged to work on meaningful products that connect people, support expression, and create value for users and communities.

Diversity and Inclusion

The company states a commitment to building an inclusive workplace where people are respected for their skills, experiences, and perspectives. Its mission to inspire creativity and enrich life is paired with a focus on reflecting the communities it serves.