Backend Engineer Graduate (TikTok Recommendation Architecture) - 2026 Start (PhD)
Core
Design and optimize the architecture of TikTok's recommendation system to ensure stability, high availability, and performance for online services and offline data flows.
Role type
Senior IC backend engineer (recommendation architecture)
Builds
High-performance storage and computing systems, data mid-platforms, and scalable recommendation infrastructure
Domain
Internet / Recommendation Systems / Large-scale Data Processing
Deliverable
production ML models
Required skills
Large model optimization, adaptive fault diagnosis, cross-domain data processing, high-concurrency real-time data writing, Schema Evolution, multimodal heterogeneous computing, data-model correlation analysis, C/C++ or Python
Preferred skills
Research in NLP/CV/data modeling, algorithm optimization, generative technology cost-efficiency
Technologies
Python, C/C++, Large Models (CV/NLP/Multimodal), DCAI
Responsibilities
Build intelligent systems for standardized recommendation strategy definition and automatic retirement; Optimize system parameters and configurations using large model capabilities; Balance costs and efficiency for generative technologies in recommendations; Handle massive heterogeneous data in cross-domain scenarios; Develop low-cost, high-performance storage engines and automated training data pipelines; Construct multimodal data heterogeneous computing frameworks; Optimize large-scale computing models for recommendation scenarios.
Seniority
Graduate (PhD), hands-on IC
