Machine Learning Engineer, Speech Synthesis
Location: London, UK
Summary & Opportunity
AJA.LA Studios is a funded early-stage startup developing speech and natural language understanding technologies for under-resourced languages. We are looking to hire an engineer, to be based in London, to participate in developing unit-selection and parametric speech synthesis for a broad library of under-resourced languages. This role provides a unique opportunity to pursue research and commercialization of speech recognition for under-resourced languages.
Ideally, candidates should be comfortable working with large quantities of data, have an interest in and/or demonstrate experience working with under-resourced languages, and an interest in working on the entire R&D/product-development cycle.
Skills & Requirements
The ideal candidate should possess a combination of the following skills and qualifications
- Masters or PhD in an analytical discipline through which you have acquired a strong knowledge of topics including
- Theory and practice of speech synthesis and/or speech processing, e.g. vocoding
- Signal Processing/Pattern Recognition
- Probability theory
- Bayesian inference
- Machine learning and related topics
- Strong software development skills
- Required: C/C++, Python, CUDA/Nsight IDE, shell scripting, Perl, Github/SVN
- Optional/Additional: Java/Android/Gradle/Android Studio, Objective C/Xcode/Cocos2dx
- Speech processing, Neural Network and Natural Language platforms and libraries
- Festival, HTK, and HTS
- Theano, PDNN, pyTorch, TensorFlow
- Operating Systems: Unix/Linux/Mac OS
We offer a compentitive salary, pension contribution, private medical insurance, and share options, flexible working hours, amongst other benefits.