ML Infra Engineer Graduate (Ads Infra) - 2027 Start
Core
Build highly efficient and stable infrastructure to collect and process data for machine learning training, serving, and privacy enforcement in TikTok's ads system.
Role type
ML Infra Engineer (Graduate)
Builds
ML platforms for distributed training, model distribution, online serving, GPU allocation, and data pipelines for sample generation and feature processing.
Domain
Advertising technology, Machine Learning Infrastructure, Distributed Systems
Deliverable
infrastructure
Required skills
C++, Python, Go, Java, data structures, algorithms, machine learning, computer systems, deep learning frameworks, distributed training, GPU scheduling, cloud-native technologies
Preferred skills
recommendation systems, MLLMs, large-scale data processing, AI agents, workflow orchestration
Responsibilities
Develop and optimize advertising recommendation models covering training, evaluation, inference, and user behavior sequence modeling; Build large-scale data pipelines for sample generation, joining, feature processing, storage, and data quality assurance; Develop ML platforms for distributed training, model distribution, online serving, and lifecycle management; Explore MLLM applications in advertising including data production, model training, fine-tuning, and inference optimization; Apply AI agents to workflow orchestration, resource scheduling, failure diagnosis, and automated operations; Build machine learning platforms for GPU allocation, scheduling, usage metering, cost accounting, and utilization optimization.
