Member of Technical Staff, Applied AI – Speech
$300k - $450k • Remote • Remote (within the U.S. only) • Full-time
Posted 9d ago
Remote Work Policy
Fully remote
Employment Type
Full-time
Categories
AI Engineer
About the job
Salient is seeking a Member of Technical Staff to join their elite Applied AI team. This role focuses on building the speech-to-speech and real-time voice models that power Salient's production voice agents, which automate various financial services processes. You will tackle complex challenges in conversational speech, including full-duplex interaction, streaming inference, and neural codecs, aiming to bridge the gap between academic benchmarks and real-world performance in regulated environments. This is a high-ownership, fully remote position where your work will directly impact live borrower conversations and contribute to raising end-to-end agent quality.
Responsibilities
- Design, train, and ship speech-to-speech and real-time voice models to enhance conversational quality on live calls.
- Build and improve the production speech stack, including neural codecs, streaming inference, ASR/TTS pathways, turn detection, and speech enhancement.
- Develop evaluation metrics and monitoring systems that accurately reflect real-world borrower conversations.
- Diagnose and resolve real-world speech failures, translating them into model, data, or systems improvements.
- Collaborate with Applied AI and Voice AI teams to enhance end-to-end agent quality.
- Take ownership of major technical areas from problem definition to production deployment.
Requirements
- 5+ years of experience in speech, audio, or related machine learning, or equivalent demonstrated depth.
- Proven experience with speech-to-speech, full-duplex, or real-time conversational speech systems.
- Public proof of work (papers, GitHub, Hugging Face models/datasets).
- Strong engineering fundamentals and ability to build reliable production systems, including streaming or low-latency inference.
- Ability to diagnose and practically improve real-world model failures.
- High ownership, strong execution speed, and comfort with ambiguity.
Benefits
- Competitive compensation
- Equity
- Work on the full stack of production voice AI
- Opportunity to own and lead core parts of the speech stack
- Ship improvements affecting millions of live conversations
- Tackle unsolved real-world speech problems
