Senior Deep Learning Researcher, ASR (PhD)

AssemblyAI is growing quickly, and we’re searching for a Deep Learning Researcher to join our team. With significant investment and strong leadership to fuel our growth, it’s the perfect time to join the AssemblyAI team! In this role you’ll have the opportunity to: Continually push the state of the art in speech recognition and NLP. You will conduct cutting edge deep learning research in speech and NLP. Our ASR models already outperform companies like Google, AWS, and Microsoft - and we’re looking for the best deep learning engineers and researchers to help us continually push the state of the art in speech recognition to get to human level performance. Research & train AssemblyAI’s deep learning models. You will work with large scale data sets to research and train deep learning models for speech recognition. You will research and develop novel algorithms and techniques to advance the state of the art in natural language understanding. You will conduct research and experiments in order to improve accuracy of our Deep Learning ASR pipeline. You will also dig into weaknesses and failure points of our current ASR models in order to identify further areas for improvement. Be part of a world-class team of creative researchers & engineers. You will help strengthen the position of AssemblyAI as a leading company in AI research. Our deep learning team is a tight knit group of creative researchers and engineers, who are not afraid to try unconventional ideas. You will also work with the broader speech recognition team as we continually strive to match human level accuracy. Requirements PhD in computer science, machine learning, or similar field 1-3+ years of professional experience in Deep Learning research & development 2+ years of experience with PyTorch or TensorFlow In-depth experience with Asynchronous (async) speech recognition Wav2Letter++, Wav2vec, CTC, RNNTs, and/or LAS Track record of publications at Speech Processing/NLP conferences (ACL, EMNLP, ICASSP, INTERSPEECH, ICLR, NeurIPS, and others) Preferred Experience (any of the following) Experience with Google JAX Experience with Reinforcement Learning Experience training large models on multiple GPUs Speaker diarization Multilingual speech recognition Real-time speech recognition Skills Detail-oriented, analytical, and creative problem solver with a passion for quality processes Ability to work independently, raise issues and take corrective action A keen eye for detail Apply tot his job Apply tot his job