Come develop audio machine learning technology with Modulate's Machine Learning Team!

Were looking for talented and flexible engineers who can employ their technical skills in speech & natural language processing, generative modeling and machine learning at Modulate. Machine learning is the backbone of our voice-native products; in addition to building the foundation of new products to enable safer and more inclusive spaces, youll be able to work on

* Expanding existing products like VoiceWear, which enables real-time voice conversion via a speech-to-speech model.

* Maintaining and improving our flagship product, ToxMod, by developing more sophisticated models to analyze voice conversations in order to better detect online harms. Combining signals from transcription, emotion, demographics, and other models evaluating multi-participant conversations, the scoring system is at the heart of ToxMod's ability to understand when harm is occurring, categorize it, and escalate it to moderators.

Modulate's machine learning team is at an exciting inflection point, defining new standards and processes to power future growth while shipping model updates and new features to a fast-growing customer base. Your skill and expertise will contribute to shaping strong research & development principles and practices for Modulate as a whole.

The salary for this position is set at $150,000/yr.

To avoid losing progress on your application, please feel free to complete responses separately in a local or cloud-saved document. If you encounter an error, any responses entered in the fields on the application will not be saved.

No unsolicited agencies, please.

Role Responsibilities:

• Train deep learning models using State-of-the-Art techniques to accomplish a variety of tasks across multiple voice-native products (GANs, diffusion models, transformers, etc)
• Build robust in-domain datasets and analyze model performance in order to deploy models to customers
• Support development of novel scoring algorithms and improve existing ones using model signals to more accurately evaluate the risk of harm in voice conversations
• Design, spec, estimate timelines for, and deliver machine learning projects/objectives
• Own (implement, improve and maintain) models important to overall company objectives
• Mentor other engineers and act as a resource capable of discussing/representing machine learning at Modulate to other members of the team

Desired Qualifications:

• Experience building deep neural networks with modern machine learning frameworks and tools, such as PyTorch or TensorFlow
• Experience with generative modeling (audio/speech modeling would be a bonus!)
• Experience with digital signal processing and/or familiarity with audio
• Experience collaborating on a codebase with a team using version control tools like git
• Experience with foundational data science concepts and practices (data collection/processing, exploratory data analysis, statistical reasoning, etc)
• Bonus: experience with deploying models to production infrastructures and optimizing those models

NOTE for the questions "Your fit for the role", "Your values/goals", and "Why Modulate?" on the following form:

Please avoid disclosing any details which would directly reveal your race, age, gender, ethnicity, sexual orientation, or other protected demographic status. We are only looking for information which directly relates to your ability to succeed in the given role. (For this same reason, resumes will be redacted before review during the initial steps of the hiring process. Its been shown that resumes sometimes lead to biases in hiring processes. If you feel that elements of your resume directly correspond to the questions below, though, feel free to copy them in.)

Additionally, we want to hear from YOU! While we understand the convenience of productivity tools and generative AI to help apply, please note that any submissions with substantial overlap or duplication with other applicant profiles will NOT be considered.