Senior AI Engineer

China Mobile International Limited; Hong Kong; On-site; HKD 600,000 - 900,000; Full time

Job summary

China Mobile International Limited is seeking a seasoned Backend/AI engineer to architect and evolve a high-performance Intelligent Speech Service Platform (ASR, TTS, NLU). The role focuses on building scalable, low-latency inference services, enabling rapid integration across diverse business scenarios.

You will collaborate with algorithm teams to productionize speech models, and contribute to cloud-native infrastructure, GPU scheduling, and large-scale auto-scaling inference clusters.

Qualifications

• Master's degree or higher in Computer Science, Artificial Intelligence, Software Engineering, or related field.
• 3+ years of backend development or AI engineering experience with hands-on background in AI algorithm platforms, model serving platforms, or speech/audio-visual systems.

Responsibilities

• Architect and drive the evolution of the Intelligent Speech Service Platform (ASR / TTS / NLU).
• Design high-concurrency, low-latency inference service architectures and microservice encapsulations.

Skills

Backend development

AI engineering

Go

C++

Python

Java

Microservice architectures

High-concurrency system design

Data structures

Cross-functional collaboration

Education

Master's degree in Computer Science/AI/Software Engineering

Tools

PyTorch

TensorFlow

TensorRT

ONNX Runtime

Triton Inference Server

Docker

Kubernetes

Job description

• Platform Architecture & Evolution: Architect and drive the evolution of the Intelligent Speech Service Platform (ASR / TTS / NLU). Design high-concurrency, low-latency inference service architectures and microservice encapsulations to enable rapid integration across diverse business scenarios.
• Algorithm Engineering & Productionization: Partner with algorithm teams to productionize speech models. Establish model versioning, automated deployment pipelines, and evaluation frameworks to ensure smooth, stable transitions from lab environments to production.
• Inference Acceleration & Performance Tuning: Optimize system architectures and call paths. Utilize model quantization, pruning, and inference frameworks (e.g., TensorRT, Triton) to boost throughput, lower response latencies, reduce GPU memory usage, and maximize computing efficiency.
• Infrastructure & Resource Scheduling: Contribute to cloud-native AI infrastructure by managing GPU compute resources, implementing Kubernetes auto-scaling, and optimizing distributed inference scheduling to improve platform delivery efficiency.
• Data Closed-Loop & Quality Evaluation: Build speech data and performance evaluation systems, streamline badcase analysis workflows, and drive data closed-loops for continuous algorithm iteration.
• Business Enablement & Agent Architecture: Contribute to the backend architecture for AI Agents and smart speech products. Design cloud-edge collaboration capabilities and multi-scenario interfaces to support rapid product iteration and large-scale deployment.
• Service Reliability & SRE: Build SLA metrics, distributed tracing, monitoring/alerting, and disaster recovery mechanisms. Drive rapid online fault diagnosis, root cause analysis, and performance optimization to guarantee 99.99% service availability.

Qualifications

• Education & Experience: Master's degree or above in Computer Science, Artificial Intelligence, Software Engineering, or related fields. 3+ years of backend development or AI engineering experience, with hands-on background in AI algorithm platforms, model serving platforms, or speech/audio-visual systems.
• Core Backend Engineering: Proficient in at least one language among Go, C++, Python, or Java. Solid fundamentals in data structures, microservice architectures (RPC, databases, caching), and high-concurrency system design.
• Model Serving & Inference Engines: Familiar with PyTorch/TensorFlow ecosystems. Hands-on experience with inference engines such as Triton Inference Server, TensorRT, and ONNX Runtime, along with proven expertise in model quantization and acceleration.
• Cloud-Native & GPU Scheduling: Familiar with Docker, Kubernetes, GPU scheduling, and cloud-native distributed architectures. Experience in building large-scale auto-scaling inference clusters is a strong plus.
• Service Reliability: Proficient in service monitoring, log tracking, performance profiling, and SRE best practices. Excellent troubleshooting skills for complex production incidents.
• Soft Skills: Strong problem-solving, cross-functional collaboration (algorithm, business, and DevOps teams), and project ownership skills.
• Competitive salary aligned with market standards (negotiable based on candidate experience and capability).
• Comprehensive benefit packages and clear career growth pathways.