Audio Machine Learning Data Engineer

Key Skills

Java Programming, Python Programming, SQL, Scala Programming

Job Description

Our Video Collaboration products help companies stay connected and can be found in the conference rooms of some of the biggest businesses in the world. We are a small nimble group and we are growing our engineering team. We create amazing experiences in the conference room using creative design, surprising science, and innovation. We love what we do because we are passionate about connecting people and we use the products we create every day to stay connected to our team.

Want to be part of this team of collaboration innovators? We are looking for a talented Audio Data Engineer to source and process the most important part of our machine learning audio algorithms; the data that drives them. You will work with machine learning data vendors, universities, and studios to source audio files to be used for training of machine learning models. You will investigate, develop, and sew together Python packages to build our pipeline of audio data pre-processing. These will be used to convert, enhance, clean, and augment the audio ready for powering the audio machine learning models.

Your Contribution:

• You will have practical knowledge of audio processing techniques such as applying filters (FIR / IIR) and convolution.

• You will have expertise in audio processing with Python, and be proficient with manipulating audio files, building functions and classes.

• You will understand audio specifications such as sample rate, bit-depth, recording environment, microphone types, background noise levels, and acoustic environment and be able to communicate these to third parties.

• You will be happy liaising and managing the work of several large companies which will record and source audio files for you.

• You will work with our Machine Learning Engineers to understand their data needs. This might include different languages of speech, less commonly available types of speech, recordings of unwanted sounds, or even acoustic soundscapes.

• You will be able to translate these audio / acoustic requirements into concrete requests to the data vendors. E.g. We may require high crest factor noise with low direct to reverberant field ratio, and you will understand and interpret this, then source and supply impulsive noises recorded at a distance in reverberant rooms.

• You will understand audio concepts such as: objective acoustic testing standards, frequency domain representations, as well as magnitude and phase of audio.

• You will have passion to work on new and disruptive AI / deep learning / machine learning solutions in the field of improving audio processing.

• You will be organised and meticulous in your record keeping of datasets investigated, and reasons for purchase / rejection. You will maintain databases or both suppliers and datasets.

Key Qualifications:

• Audio expertise, understanding of digital audio processing.

• Acoustics expertise, understanding of acoustic concepts such as RT60, Clarity, STI, DRR.

• Audio Recording Experience, knowledge of recording techniques, audio hardware, and microphone types.

• Expertise in audio processing with Python.

• Understanding of basic machine learning concepts. Awareness of and interest in machine learning and the power of big data.

• People skills, managing relationships with data suppliers.

• Organisation and record keeping.

• Project management, ability to manage several projects in parallel.

• Understanding of open source audio data repositories, such as TIMIT and MUSAN.

• Preferred Qualifications:

  • Experience / awareness and understanding of machine learning.

  • Experience in using Source Control / Project Tracking systems (Github, Jira).

  • Basic Linux / Unix terminal expertise.

  • C++ / Matlab experience is a bonus but not required.

  • Contribution to any open-source programming or audio processing tools.

• Education:

  • BSc, MSc, or PhD in Audio Engineering, Acoustics, or equivalent expertise.