Member of Technical Staff - Multi-Modal, Audio

What you'll do

• Build and scale audio-model training data pipelines, including preprocessing, augmentation, and quality filtering.
• Design and maintain evaluation systems for multimodal performance across internal and public benchmarks.
• Fine-tune audio models for customer use cases and deliver them through deployment; contribute production code to the core audio repository.

What they ask for

• Strong programming fundamentals and ability to write clean, maintainable, production-grade code.
• Experience building and shipping production ML systems beyond model training, such as data pipelines, evaluations, or serving infrastructure.
• Proficiency in PyTorch and familiarity with distributed training frameworks such as DeepSpeed or FSDP.
• Track record of collaborating effectively in shared codebases with high engineering standards.

Nice to have

• Direct experience with audio or speech models, such as ASR, TTS, vocoders, diarization, or speech-to-speech systems.
• Experience with large-scale distributed-GPU training experiments or relevant open-source contributions.

Tools

PyTorch, DeepSpeed, FSDP

AI focus

Develop and deploy speech-language models for STT, TTS, and speech-to-speech, including customer-specific audio-model adaptation and evaluation for on-device use.