AI Engineer, Cantonese Speech
Hong Kong Applied Science and Technology Research Institute
- Location
- Hong Kong
- Level
- Mid
- Posted
- 6 Aug 2026
MidOn-siteActive
Original posting
Skills
- Speech recognition
- Cantonese
- ASR
- PyTorch
- Signal processing
Summary
ASTRI is hiring an AI engineer for Cantonese and mixed Cantonese-English speech systems used by public-sector partners. The work is applied speech: acoustic models, code-switching, and noisy far-field audio from real counters. You will fine-tune recognition models, build test sets from consented recordings, and package a model the solutions team can deploy on premise. The office is in Hong Kong. Requirements: speech or audio machine-learning experience, Python, and enough Cantonese to judge transcription errors. A research degree is optional.
Full description
ASTRI is hiring an AI engineer for Cantonese and mixed Cantonese-English speech systems used by public-sector partners. The work is applied speech: acoustic models, code-switching, and noisy far-field audio from real counters.
You will fine-tune recognition models, build test sets from consented recordings, and package a model the solutions team can deploy on premise. The office is in Hong Kong.
Requirements: speech or audio machine-learning experience, Python, and enough Cantonese to judge transcription errors. A research degree is optional.