Speech Data scientist

Speech Data Scientist –

We're looking for Speech Data Scientists, to develop high-impact, high-visibility Speech AI product "Riva" & improve the experience of millions of customers.

While the overall profile should still be a person with TTS model training experience, instead of looking for SOTA models, Looking for people more focused on speech data sourcing, clean-up, processing, and evaluation?

What you’ll be doing:

• Train Speech Synthesis Mel spectrogram and vocoder models.
• Measure and benchmark model performance.
• Maintain TTS model evaluation system.
• Analyze model accuracy and bias and recommend the next course of action & Improvements.
• Improve processes for speech data processing, augmentation, filtering & TTS Training sets preparation.
• Gather know how on TTS datasets for training & evaluation.
• Characterize performance and quality metrics across platforms for various speech AI components. Collaborate with various teams on new product features and improvements of existing products.
• Participate in developing and reviewing code, design documents, use case reviews, and test plan reviews.
• Help innovate, identify problems, recommend solutions and perform triage in a collaborative team environment.

What we need to see:

• Master’s degree (or equivalent experience) or PhD in Computer Science, Electrical Engineering, Artificial Intelligence, Applied Math, Linguistics or Computational Linguistics
• 5+ years of experience
• Excellent programming skills in Python.
• Strong fundamentals in Programming, optimizations and Software design.
• Strong knowledge of ML/DL techniques, algorithms and tools with exposure to CNN, RNN (LSTM), Transformers.
• Knowhow of Deep learning applications to Speech synthesis, LLM, and Speech-to-speech translations.
• Hands-on experience on Speech Technologies like Speech Synthesis, voice cloning, etc.
• Experience with Training of speech models.
• Experience with “PyTorch” Deep Learning Frameworks.
• Exposure to basic speech digital signal processing and feature extraction techniques like FFT, MFCC, Mel Spectrogram, etc.
• General background around version control and code review tools like Git, Gerrit, Gitlab.
• Strong collaborative and interpersonal skills, specifically a proven ability to effectively guide and influence within a dynamic matrix environment.

Ways to stand out from the crowd:

• Native or near-native fluency in a non-English language - Spanish / Mandarin / German / Japanese / Russian / French / UK English / Arabic / Hindi / Korean / Italian / Portuguese.
• Experience developing multilingual code-switched TTS, voice cloning, and cross-lingual voice cloning.
• Experience developing WFST and Neural networks-based Text-Normalization and Inverse Text-Normalization.
• Experience working with G2P systems for multiple languages.
• Strong personal interest in learning, researching, and creating new technologies related to foreign languages, linguistics, phonetics, phonology and language technology.
• Feeling comfortable and motivated when working in a fast paced, highly collaborative, dynamic work environment.
• Strong C++ programming skills.
• Familiarity with GPU based technologies like CUDA, CuDNN and TensorRT
• Background with deploying machine learning models on data center, cloud, and embedded systems.