AI Infrastructure Engineer Intern (TikTok Recommendation Architecture) - 2027 Start
Core
Design, build, and optimize distributed training and low-latency inference systems for TikTok's recommendation models and Large Language Models (LLMs).
Role type
AI Infrastructure Engineer Intern
Builds
Distributed training infrastructure, online inference systems, and GPU optimization for recommendation and LLM workloads
Domain
AI Infrastructure, Large-Scale Distributed Systems, GPU Computing
Deliverable
production ML models
Required skills
C++, Python, PyTorch, TensorFlow, Transformer architectures, data structures and algorithms
Preferred skills
LLM training/inference experience, distributed training concepts (DP, TP, PP, FSDP, ZeRO), GPU programming (CUDA, Triton)
Technologies
PyTorch, TensorFlow, CUDA, Triton, FSDP, ZeRO
Responsibilities
Design and optimize distributed training infrastructure; Develop high-performance GPU kernel implementations; Build compiler optimization passes and operator fusion technologies
Seniority
Intern