On-Device AI Engineer: Speech Models & Tooling

Shokz; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

Shokz is seeking an AI Engineer to help build and optimize speech models (ASR, TTS) and pipelines for wearable devices. You will work across requirements, design, debugging, and iteration to improve scenario performance and response stability.

Responsibilities include on-device deployment, quantization, and inference optimization, plus collaboration with product, algorithm, system and testing teams to drive data-driven improvements and robust experiences.

Qualifications

• Solid understanding of DL architectures (Transformer/Conformer/BERT) and ASR/LLM tech basics.
• Strong programming skills in Python, C++, or Java.
• Experience with model deployment and on-device inference concepts.
• Familiarity with AI app development approaches (RAG, prompts, orchestration).
• Knowledge of ONNX, TFLite, or similar frameworks is a plus.
• Able to diagnose common performance, latency and reliability issues.

Responsibilities

• Contribute to building and optimizing speech models (ASR, TTS) and pipelines.
• Participate in end-to-end AI feature development for wearables, including design and testing.
• Engage in on-device model deployment, quantization, and performance tuning.
• Support AI agents with capabilities like task planning, tool invocation, and memory.
• Collaborate with product, algorithm, system, and test teams on data analysis and evaluation.
• Contribute to engineering best practices including logging, release management, and stability.

Skills

Deep learning

Python

C++

Java

RAG

On-device AI

Education

Bachelor’s degree or above in Computer Science / AI / Software Engineering

Tools

ONNX

TFLite

NCNN

MNN

ONNXRuntime

Job description

Shokz is seeking an AI Engineer to help build and optimize speech models (ASR, TTS) and pipelines for wearable devices. You will work across requirements, design, debugging, and iteration to improve scenario performance and response stability.

Responsibilities include on-device deployment, quantization, and inference optimization, plus collaboration with product, algorithm, system and testing teams to drive data-driven improvements and robust experiences.