Temasek Laboratories @ NTU invites applications for the position of Research Assistant.
The project studies algorithms for low resource automatic speech recognition (ASR) under cross-domain conditions for real-world applications. This is a challenging problem which is receiving much attention in the ASR research community recently, and research in low-resource and cross-domain is an on-going issue in the ASR research community.
The candidate will work to meet project requirements, specifically focusing on the low-resource and cross-domain ASR algorithms research.
- Data preparation of text and acoustic data to train ASR system
- Lexicon modeling, adding new words and corresponding pronunciation to the lexicon
- Language modeling, update language models, and state-of-the-art neural network language modeling
- Training state-of-the-art ASR system using Kaldi and new end-to-end toolkit
- Bachelor's degree in computer science/engineering or related fields
- At least one-year ASR engineer experience on training speech recognition models including acoustic models, language models, and End-to-End ASR models
- Experience with Speech Recognition and AI frameworks including KALDI, tensorflow, and pytorch
- Possessing research experience and publications in ASR and its related fields is preferable
- Strong in programming languages, i.e. C/C++, scripting languages as python, perl
- Strong experience in working with Linux OS environment and high-performance computing cluster
- Familiar with Linux bash/shell packages such as awk, sed
We regret that only shortlisted candidates will be notified.