Machine Learning Research Scientist

Key Skills

Java Programming, Python Programming, C++, C Programming, SQL, Cloud computing, Scala Programming, Machine learning techniques, Data science techniques, MATLAB Programming, PyTorch, TensorFlow, R Programming

Job Description

Wispr is building the future of human-computer interfaces controlled by the peripheral nervous system. We're building a team of world class scientists, engineers, and product designers to make that a reality. We're funded by top tier Silicon Valley investors (NEA and 8VC) and backed by angels including Ben Jones (COO, CTRL-Labs), Jose Carmena (Berkeley professor; co-CEO iota), David Gilboa (CEO, Warby Parker), and Will Ahmed (CEO, Whoop). Our founders are Stanford alums and have previously sold a company and run a team at a deep tech startup with over 100M in funding.

We're hiring in SF Bay Area / Remote.

As an applied ML research scientist at Wispr, your work closely with neuroscientists and machine learning engineers to design algorithms to decode surface electromyography (EMG) and other neural signals to speech. You will be responsible for designing experiments for scaled data collection, investigating deep learning architectures for speech processing, and inventing novel techniques for extracting signal from EMG (unsupervised pretraining, semi-supervised learning, synthetic data, multi-modal feature extraction, among others).

Come join us and make magic happen.

Core Job Responsibilities

• Collaborate with neuroscientists, machine learning engineers, and software engineers to build speech decoding algorithms from neural signals
• Determine best ways to scale data collection and leverage large datasets
• Design and iterate on deep learning architectures for speech recognition on novel sensors, trade off between different designs for transcription models
• Invent and adapt novel techniques for extracting signal from EMG (unsupervised pretraining, semi-supervised learning, synthetic data, multi-modal feature extraction, among others).
• Setup best practices for ML research model versioning

Required Knowledge/Skills, Education, And Experience

• PhD in computer science, machine learning or related engineering field, or equivalent research experience
• 3+ years hands-on relevant research experience in speech processing (ASR) or natural language processing
• 2+ years industry experience building speech recognition systems from scratch, including exposure to best practices for model and dataset versioning
• Strong research/publication track record
• Experience in Python and machine learning and deep learning libraries (Pytorch, NumPy, Pandas, etc)
• Experience with cloud computing (AWS, Google Cloud)
• Organized, self-directed, efficient and able to manage priorities and expectations
• Experience working collaboratively with teams: we believe in working collectively towards a common goal

Nice to have Knowledge/Skills, Education, And Experience

• Experience working with bio-signals (EMG, EEG, EKG, etc), medical devices, or real world noisy data
• Experience with novel methods for increasing performance of speech recognition systems, including leveraging synthetic data, unsupervised pretraining, etc.