AI Engineer, Agentic Voice at Dialpad. About the role You will lead the backend development and linguistic refinement of the voice synthesis layer for our next-generation AI agents. This position focuses on bridging speech science and product engineering to ensure our agents sound natural, context-aware, and reliable.
Key facts Location: Vancouver, Canada
Engagement: Hybrid
Team: Speech Team
What you'll do Integrate and optimize various TTS vendor APIs behind a unified interface with failover capabilities.
Conduct research and prototyping for open-source and internal TTS architectures.
Minimize end-to-end latency across the ASR, LLM, and TTS pipeline while maintaining audio quality.
Apply phonetics and sociolinguistics to improve TTS output using SSML tags, prosody, and normalization for dates, numbers, and currency.
Build persona parameterization systems and architect logic to expose voice attributes to the UI.
Manage versioned LLM and TTS prompt templates using a structured evaluation harness.
Design conversational utterances that maintain user trust during background model processing.
Requirements 3+ years of experience in Speech Synthesis (TTS) or applied speech machine learning.




Proficiency in Python and deep learning frameworks like PyTorch.
Hands-on experience with tools such as NVIDIA NeMo, ESPnet, or Coqui.
Experience with major TTS APIs including ElevenLabs, Rime, and Cartesia.
Degree in Computer Science, Computational Linguistics, or AI/ML.
Understanding of phonetics, prosody, and syntax, including G2P and pronunciation dictionaries.
Experience building production-grade APIs and integrating multi-vendor services in cloud environments (GCP preferred).
Ability to design A/B tests for voice personas using metrics like MOS, intelligibility, and latency.
Proven experience in crafting and versioning LLM prompts.
Skills & tools Python, PyTorch
NVIDIA NeMo, ESPnet, Coqui
ElevenLabs, Rime, Cartesia
SSML, G2P
GCP
LLM prompting
Practical notes Salary range for British Columbia: $161,500 to $191,500 CAD base salary.
Compensation excludes bonus, equity, and advantages.
This is a production engineering role; candidates focused solely on conversation design or prompt-tuning should apply to the Lead Voice Experience Designer role instead.

#J-18808-Ljbffr

📌 AI Engineer, Agentic Voice (British Columbia)
🏢 Dialpad
📍 British Columbia