ASR Engineer

About this role

Join an early-stage ambient intelligence hardware startup as an ASR Engineer to own the end-to-end transcription pipeline, balancing cloud-based systems with minimal on-device components. You'll ship production improvements across latency, accuracy, and reliability while collaborating directly with product leadership and international R&D teams.

What you'll do

• Build and iterate on cloud-based ASR pipeline from audio capture through post-processing at production scale
• Own ASR quality metrics end-to-end, delivering measurable improvements to latency, small-word accuracy, and voice-print reliability
• Execute full ML lifecycle: data preparation, model training/fine-tuning, evaluation, and deployment based on product feedback
• Collaborate across distributed teams including China-based R&D, hardware, and supply-chain groups
• Partner with product engineers on shared backend and pipeline infrastructure
• Translate loosely-defined product requirements into concrete, production-ready pipeline changes

What they're looking for

• ASR/transcription pipeline development and tuning in production environments
• Cloud-based audio processing and streaming pipeline optimization
• ML model training, fine-tuning, and evaluation frameworks
• Latency-sensitive system design and optimization
• On-device ML frameworks (Core ML, TensorFlow Lite, or equivalents)
• Production ML systems deployment and monitoring
• Audio/signal processing fundamentals
• Asynchronous cross-timezone collaboration

Benefits

• Early-stage startup equity and founding-stage impact
• 3 days per week hybrid work in San Francisco Bay Area
• Direct collaboration with GM and Head of Product
• Ownership of critical production system end-to-end