High-Performance Backend Inference Engineer

Bytedance; Singapore; On-site; SGD 180,000 - 280,000; Full time

Job summary

ByteDance in Singapore is seeking a skilled engineer to lead the architecture design and implementation of model inference services for large-parameter AI models. You will build a high-performance, scalable inference system and drive optimization across engineering teams.

The role covers research and development of core modules, including inference engine scheduling, monitoring, canary releases, and performance tuning in high-concurrency environments.

Qualifications

• Bachelor's or Master’s in computing or related discipline.
• Solid C/C++ programming skills and knowledge of data structures and algorithms.
• Familiar with Linux commands and multi-threaded concurrency concepts.
• Experience in R&D of high-concurrency distributed services and understanding latency/resource optimization.
• Strong learning, problem-solving, and cross-team collaboration abilities.

Responsibilities

• Design and implement the architecture of model inference services for large-parameter AI models.
• Develop core modules of the inference framework, including scheduling, monitoring, and canary release.
• Stay updated on industry inference technologies and drive performance and standardization improvements.

Skills

C/C++ programming

Linux

Multi-threading

Data structures & algorithms

Concurrency

Performance tuning

R&D high-concurrency distributed

Problem solving

Education

Bachelor's or Master’s in computing or related

Tools

Redis

RocksDB

BRPC

GRPC

GPUs

Job description

ByteDance in Singapore is seeking a skilled engineer to lead the architecture design and implementation of model inference services for large-parameter AI models. You will build a high-performance, scalable inference system and drive optimization across engineering teams.

The role covers research and development of core modules, including inference engine scheduling, monitoring, canary releases, and performance tuning in high-concurrency environments.