PhD Infrastructure Intern – AI/ML Infrastructure

Job Description

Student Researcher (AI Foundation Models Infrastructure – Seed Infra) – 2026 Start (PhD) | ByteDance

The Tone:
ByteDance’s Seed Infrastructures team builds systems for AI foundation models, focusing on distributed training, inference, and hardware compilation. This internship allows students to contribute to products, research, and emerging technologies. The work supports advancing artificial general intelligence, driving progress for technology and society.

The TL;DR
• Role: Student Researcher (AI Foundation Models Infrastructure – Seed Infra) – 2026 Start (PhD)
• Type: Internship
• Team: The Seed Infrastructures team oversees the distributed training, reinforcement learning framework, high-performance inference, and heterogeneous hardware compilation technologies for AI foundation models.
• Mission: Actively contribute to products, research, and emerging technologies to advance artificial general intelligence.
• Tech Stack: PyTorch FSDP, Megatron-style parallelism, CUDA, Triton, Python, C++

What You’ll Actually Do
• Distributed Training: Design and optimize large-scale distributed training systems for efficiency and fault tolerance.
• Reinforcement Learning: Contribute to training frameworks and large-scale post-training systems.
• Inference Performance: Improve latency and throughput for foundation models.
• Hardware Optimization: Develop compiler or runtime optimizations for heterogeneous hardware.

The Must-Haves
• Background: PhD degree in Computer Science, Electrical Engineering, or related fields with a solid understanding of systems, distributed computing, or ML systems.
• Experience: Background with distributed training frameworks, reinforcement learning systems, GPU programming, or large-scale inference optimization; experience on large-scale ML systems preferred.
• Skills: Strong programming skills in Python and/or C++.