ML Infra Engineer: Real-Time Recommendation & Search

TikTok Inc.; Seattle (WA); On-site; USD 130,000 - 246,000; Full time

Benefits offered by this job

Medical, dental, vision insurance

401(k) with company match

Paid parental leave

Job summary

TikTok USDS Joint Venture LLC is seeking highly skilled ML engineers to advance recommendations and search with large-scale distributed training and real-time inference platforms. You will own system design, parallelism, and low-latency serving for multi-thousand GPU clusters.

The role emphasizes collaboration with applied ML researchers and data platforms, working in a fully in-person environment in Seattle.

Qualifications

• Bachelor’s or Master’s in Computer Science, Computer Engineering, or related technical discipline.
• 2+ years of professional software engineering with Python, C++/Java.
• 1+ years building and maintaining ML infrastructure at enterprise scale.
• Deep familiarity with core ML frameworks (PyTorch/TensorFlow) and low-level GPU memory management / CUDA.

Responsibilities

• Contribute to technical roadmap and implement large-scale distributed real-time ML training and inference platforms.
• Design and scale multi-node distributed training systems with 3D parallelism (Data, Tensor, Pipeline).
• Build and scale low-latency, high-throughput model serving infrastructure.
• Partner with ML Research teams to co-design generative recommendation systems.
• Develop fault-detection and checkpointing for thousands of GPUs.
• Collaborate with data platform teams on secure, high-throughput data processing.

Skills

Python

C++/Java

ML infra experience

GPU/CUDA knowledge

Networking basics

Troubleshooting

Education

Bachelor’s or Master’s in CS/CE

Tools

PyTorch

TensorFlow

Megatron

DeepSpeed

CUDA

InfiniBand/RoCE

Job description

TikTok USDS Joint Venture LLC is seeking highly skilled ML engineers to advance recommendations and search with large-scale distributed training and real-time inference platforms. You will own system design, parallelism, and low-latency serving for multi-thousand GPU clusters.

The role emphasizes collaboration with applied ML researchers and data platforms, working in a fully in-person environment in Seattle.