• University degree in Speech/Computer Science/Engineering (Master/PhD)
• Experience with GMM-HMM and DNN based models.
• Experience with Linux environments and scripting languages (Python, Bash).
• Experience with Kaldi speech recognition toolkit.
• Experience with ASR pipelines and Weighted Finite-State Transducers (WFSTs)
• Good knowledge of advanced ASR concepts (e.g. speaker adaptation, diarization and Vocal Tract Length Normalization).
You like a challenge and are a pleasant person to work with.
Due to the distributed nature of our team, being a self starter is important.
Also communication skills and proficient English.
Having an affinity with our mission in with improving education and healthcare is important. We’d like to work with other driven and enthusiastic people.
• Working on cutting edge ASR technologies, modelling techniques and ASR algorithms.
• Contribute to ASR model optimization for in depth phonetic speech analysis.
• Develop model training pipelines and deploy production ASR models.
• Design and conduct experiments to evaluate the quality of speaker recognition models for different target groups.
• Collaborate with the software engineering team to improve data pipelines and tooling for continuously training and benchmarking ASR models.
Nice-to-have:
• Experience with CI/CD pipelines.
• Professional coding experience (C++ / Python)
• Experience with machine learning toolkits (PyTorch / TensorFlow)
• A challenge in applying machine learning in the domain of speech.
• A small team of highly skilled individuals, with a minimum amount of meetings.
• Competitive salary, profit sharing.
• Remote work and flexibility (our team has always been distributed).
• A way of applying your expertise in a way that matters to society.