Senior Software Engineer, E-Commerce AIGC Large Model Engineering - USDS

Responsibilities

About the Team:
The team is part of the e-commerce business unit under ByteDance, focusing on the R&D and platform-based implementation of AI large model technologies. It is committed to driving the intelligent upgrading of e-commerce scenarios through multimodal large models, large language models (LLM), and AIGC technologies.
Leveraging the massive user base accumulated across multiple platforms and the rapid development of e-commerce businesses, the team integrates algorithmic innovation with engineering capabilities in depth. It provides technical support for scenarios such as content understanding, intelligent customer service, platform governance, logistics, supply chain, and search & recommendation. Additionally, it empowers R&D efficiency across all business lines through platform-based tools.

In order to enhance collaboration and cross-functional partnerships, among other things, at this time, our organization follows a hybrid work schedule that requires employees to work in the office 3 days a week, or as directed by their manager/department. We regularly review our hybrid work model, and the specific requirements may change at any time.

Responsibilities:
- Design and implement a general large model inference engine for e-commerce scenarios, applicable to online real-time inference, near-offline batch inference and other scenarios, and promote its application across various e-commerce business scenarios.
- Conduct targeted inference performance optimization for multiple models such as LLM/MLLM/AIGC in e-commerce scenarios, overcoming technical challenges including high concurrency, high reliability, and high scalability.
- Responsible for the development of large model/AIGC service chains and application platforms, supporting the production and efficient iteration of related businesses.
- Build capabilities for the management, annotation, preprocessing, and storage of large-scale sample data, providing infrastructure support for model training and inference.
- Manage and optimize GPU resource scheduling for the U.S. region of international e-commerce, build management tool platforms, improve GPU management efficiency and the overall utilization rate of the resource pool, and support large model business scenarios in the U.S. region of international e-commerce.
- Explore cutting-edge deep learning technologies in the industry, continuously enhance platform capabilities, and reduce the R&D and algorithm usage costs.

Qualifications

Minimum Qualifications:
- Have practical experience with various models such as LLM/VLM/AIGC, Intelligent Agent technology, RAG technology, and AutoPrompt technology.
- Possess basic understanding of large model deployment processes.
- Proficient in Python/Go programming languages, with good programming practices and code management capabilities.
- Participate in the development and optimization of large model automatic evaluation technologies to improve model usability and performance.
- Familiar with backend fundamental technologies including but not limited to Ray, message queues, and Elasticsearch (ES), capable of efficiently developing and maintaining backend services.

Preferred Qualifications:
- Understand and be able to use common deep learning frameworks such as PyTorch and vLLM for model inference and deployment.
- Have good communication skills, able to effectively communicate with business teams, understand business requirements, and translate them into technical implementations.

Job Information

[For Pay Transparency] Compensation Description (annually)

The base salary range for this position in the selected city is $177688 - $341734 annually.