后端研发工程师-Pangle Infra
Core
Design and implement architecture for Pangle ad algorithm services (recall, coarse ranking, fine ranking) to support tens of millions of ad candidates and improve online inference performance.
Role type
Senior IC backend engineer (ad algorithm infrastructure)
Builds
High-performance ad ranking systems, deep learning model training and inference pipelines, and distributed offline/online separation architectures.
Domain
Online advertising, distributed systems, deep learning engineering
Deliverable
production ML models
Required skills
C/C++ programming, Linux environment expertise, data structures and algorithms, system design, distributed system architecture, GPU programming (CUDA/Triton), deep learning frameworks (TensorFlow/PyTorch), feature engineering, offline sample processing
Preferred skills
GPU performance analysis, search/ad/recommendation algorithm knowledge, NLP model optimization, experience with TensorFlow Serving or PyTorch distributed training
Responsibilities
Architect and implement ad algorithm services for recall and ranking; build deep learning capabilities for real-time training and inference; optimize system capacity and distributed architecture; upgrade feature engineering and offline sample production services; optimize system costs and stability under resource constraints
