Skip to content
NEXTMOVEFDE careers · United States
Together AI

Staff Machine Learning Engineer, Voice AI

AI summary of the role

Staff ML Engineer to own the model serving layer for Together AI's real-time voice platform, optimizing STT/TTS and speech-to-speech inference on H100/H200/B200 GPUs.

What you’ll do

  • Own the voice inference roadmap end-to-end, defining technical strategy for STT, TTS, and speech-to-speech optimization.
  • Architect and implement systems targeting leading TTFB, throughput, and GPU utilization for voice workloads.
  • Design serving architecture for serverless and dedicated endpoints, including batching, streaming pipelines, and memory management.
  • Build a rigorous evaluation framework covering WER, naturalness, latency, and pronunciation fidelity.

What you’ll bring

  • 8+ years ML engineering with focus on model serving, inference optimization, or ML infrastructure at production scale.
  • Deep expertise in LLM serving engines (vLLM, SGLang, TensorRT-LLM) with ability to modify internals.
  • Expert-level Python and PyTorch, strong GPU optimization skills (CUDA kernels, profiling).
  • Proven system design judgment and technical leadership in ambiguous environments.

Technologies

TRT-LLM · SGLang · vLLM · PyTorch · CUDA · Whisper · Parakeet · Orpheus · Kokoro · SNAC · Encodec · DAC

About Together AI

GPU cloud and inference platform optimized for open-source LLMs, serving 450K+ developers and enterprises with serverless APIs, dedicated clusters, and fine-tuning.

Series B

Source and classification

Deployment team leadership · Evidence for this classification:

About the Role Together AI is building the best inference infrastructure for voice applications. Our Voice AI platform powers production-grade, real-time voice agents and applications — serving speech-to-text and text-to-speech models with best-in-class latency and reliability. We're looking for a Staff ML Engineer to drive the model serving layer for voice workloads. You'll work hands-on with inference engines like TRT-LLM and SGLang to optimize how we serve models like Whisper, Parakeet, Orpheus, and Kokoro — pushing latency and throughput to the frontier. You'll profile GPU utilization, design batching strategies for streaming audio, and ensure new model architectures can go from research to production quickly. This is a foundational hire on a small, high-impact team. Voice inference has unique challenges — streaming audio, tokenization, real-time latency budgets — that require
How jobs are selected

Employer postings · Data from · Sources