Decagon
Research Engineer, Audio and Speech
$200K–$400K
San Francisco +1 · Full-time · Posted Sep 4, 2026
- Location
- San Francisco
New York City - Experience
- 2+ years of experience in speech, audio ML, multimodal ML, or production machine learning.
About the job
Build and deploy models and agent systems for real-time voice agents, advancing speech understanding, generation, and natural interaction.
What you'll do
- Design streaming speech agent harnesses for turn-taking, interruptions, overlapping speech, and continuous interaction.
- Research and train multimodal and full-duplex models; improve speech recognition, voice activity detection, endpointing, and speech generation.
- Evaluate production calls and optimize inference for accuracy, latency, naturalness, throughput, stability, and cost.
What they ask for
- Experience developing or adapting autoregressive, diffusion, flow-matching, or codec-based speech models.
- Hands-on experience with streaming agent systems, low-latency inference, production model serving, and evaluation on real-world audio.
- Fluency in Python and a modern deep-learning framework such as PyTorch, with strong foundations in machine learning and signal processing.
- Track record of taking research ideas from prototype to reliable, measurable production impact.
Nice to have
- Familiarity with speech-to-speech or full-duplex models.
- Experience with telephony, multilingual speech, noisy-channel robustness, speaker adaptation, or expressive speech generation.
Pay
$200K – $400K (USD, base salary)
Tools
Python, PyTorch
AI focus
Research, train, and deploy multimodal and full-duplex speech models and real-time voice-agent systems.
Model-written summary · the original posting is authoritative