Senior Deep Learning Data Scientist, Speech

Widely regarded as one of the most sought-after employers in the tech industry, NVIDIA leads the pack with its innovative progressions in High-Performance Computing, Artificial Intelligence and Visualization. The GPU, our invention, acts as the visual cortex of contemporary computers and is crucial to our products and services. GPU deep learning has sparked the next age of computing — modern AI — with the GPU functioning as the brains of computers, robots, self-driving cars, and AI that can comprehend and interpret the world. Nowadays, we are increasingly recognized as “the AI computing company”. We aim to expand our company and assemble our crews with the brightest minds around the globe. Join us at the vanguard of technological progression. NVIDIA is on the hunt for Speech Data Scientists to enhance the high-impact, high-visibility Speech AI product "Riva" & improve the experience of millions of users. If you're innovative & passionate about solving real life conversational AI challenges, don't hesitate to join our Riva Product engineering team. You can find more details on Riva at https://developer.nvidia.com/riva What will you do: Train Speech Recognition Acoustic, Language, Punctuation models. Measure and benchmark model performance. Maintain the ASR model evaluation system. Analyze model precision and bias, suggesting changes & Improvement. Enhance procedures for speech data processing, augmentation, filtering & ASR Training sets assembly. Learn about speech datasets for training & evaluation. Specify performance and quality metrics across platforms for various speech AI components. Work together with distinct teams on novel product features and amelioration of existing products. Take part in building and reviewing code, design documents, usage case assessments, and test plan evaluations. Help pioneer, identify issues, suggest solutions and carry out triage in a team environment. What we wish to see: A Bachelor's degree, Master’s degree or PhD (or equivalent experience) in Computer Science, Electrical Engineering, Artificial Intelligence, or Applied Math. Native or near-native fluency in a non-English language - Spanish/ Mandarin/ German/ Japanese/ Russian/ French/ UK English/ Arabic/ Hindi/ Korean/ Italian/ Portuguese Excellent Python programming skills as well as strong understandings of Programming, optimizations, and Software design. Solid comprehension of ML/DL strategies, algorithms, and tools as well as exposure to CNN, RNN (LSTM), Transformers. Good knowledge of RNNT and CTC decoders. Familiarity with the application of Deep learning to Speech and NLP. Practical experience on Speech Technologies like Automatic Speech Recognition, Speech Command detection, Text to Speech, Speaker Recognition and Identification, speaker diarization, Noise robustness Techniques, Voice activity detection, End of utterance detection, etc. Experience with Training acoustic models. Experience with KenLM, OpenLM and other tools to create Language 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 knowledge of version control and code review tools like Git, Gerrit, Gitlab. Ways to stand out: Strong C++ programming skills. Awareness of GPU based technologies like CUDA, CuDNN and TensorRT. Background in Dockers and Kubernetes. Experience in deploying machine learning models on data center, cloud, and embedded systems. NVIDIA is devoted to fostering a diverse working environment and is proud to be an equal opportunity employer. As we highly value diversity among our existing and future workforce, we practice non-discriminatory policies (including in our hiring and promotion practices) based on race, religion, color, national origin, gender, gender expression, sexual orientation, age, marital status, veteran status, disability status or any other characteristic protected by law.