Senior Research Scientist/AI Software Engineer, LLM/Agent Platform (TikTok-Content Ecology AI Innovation & Platform) - TikTok Content Ecology

The Tik Tok Content Ecology team builds the core algorithm systems for content understanding and ecosystem governance on Tik Tok. Drawing on platform-scale video content and interaction signals, we apply multimodal large models, natural language processing, computer vision, and recommendation systems to build a content foundation platform supporting search, local services, product infrastructure, PGC, and UGC across Tik Tok's global business.

Our work spans three areas:

1. Unified content understanding. We use multimodal large models for fine-grained semantic parsing of video, building a cross-category, multilingual labeling and event knowledge system that provides reliable foundations for search, recommendation, and safety.
2. Platform-level content governance. We develop high-precision recognition and review models that handle billions of pieces of content daily, addressing low-quality and non-compliant material and supporting compliance requirements for goods and content in areas such as local services.
3. Business growth and user experience. Through trending discovery, inspiration recommendations, and intelligent customer service, we serve creators and users worldwide, improving distribution efficiency, the precision of creative recommendations, and the speed of customer service response.

We work on understanding hundreds of languages and processing multimodal data at hundred-billion scale, in a global business that keeps growing. You will apply NLP, multimodal, and large model technology at real-world scale, seeing your work reach hundreds of millions of users.

The Content Ecology Algorithm Team drives Tik Tok’s AI innovations in LLMs, NLP, Computer Vision (CV), multimodal learning, and recommendation algorithms. We develop cutting-edge AI capabilities that power multiple business lines.

We are seeking exceptional, experienced Research Scientists / Architects in the areas of LLM / Agent Platform. This is a unique role for innovators who are passionate about building the "bones" (scalable infrastructure) of next-generation AI and Agent. You will design and implement novel LLM/Agent frameworks with self-improving capabilities and architect the robust, high-performance infrastructure that brings these LLMs and agents to life at a global scale. This is an opportunity to shape the future of AI at one of the world's most dynamic technology companies.

Responsibilities:

• Architect and build standardized, configurable, and reusable pipelines for the entire lifecycle of models and agents—from data processing and training to deployment, monitoring, and governance.
• Partner with algorithm teams to understand their needs and provide a world-class infrastructure platform that accelerates their research and development cycles.
• Build robust observability and evaluation frameworks to ensure the reproducibility, reliability, and cost-efficiency of AI workloads at scale.
• Design and implement core platform infrastructure, including model/agent registries, feature stores, and high-throughput retrieval/RAG systems.
• Design, build, and optimize advanced Agentic AI systems, focusing on core components like planning, tool use, and memory.

Requirements:

Minimum Qualifications:

• BS/BA or Master in Computer Science or related technical field or equivalent technical experience
• 5+ years of hands-on experience in software engineering, with a focus on machine learning, distributed systems, or AI infrastructure.
• Strong proficiency in integrating AI tools into knowledge discovery and research workflows.
• Familiarity with building robust evaluation frameworks and ensuring experimental reproducibility.
• Expertise in deep learning frameworks and tensor libraries like Py Torch, Tensorflow, JAX/FLAX
• Solid understanding of machine learning fundamentals and the modern AI stack.
• Excellent communication skills to collaborate across teams.

Preferred Qualifications:

• PhD in Computer Science or related technical discipline.
• 5+ years of experience as an architect, or technical leadership position
• Experience with the ML infrastructure ecosystem, including GPU scheduling, model serving (Triton, TensorRT-LLM), vector databases (FAISS, Milvus), and MLOps principles.
• Experience with large-scale model training and inference, including distributed training, KV cache–aware serving, GPU/accelerator optimization, and high-performance networking (e.g., RDMA, NCCL).
• Experience with performance optimization of large model training and inference (e.g., Deep Speed/ZeRO, vLLM).
• Deep knowledge of agent architectures, including planning, tool use (e.g., Lang Chain, Llama Index), and memory systems.
• Publications in systems and/or machine learning conferences (e.g., NeurIPS, OSDI, SOSP, ASPLOS, MLSys).