← Back
AWS ML Blog

PolyAI Dialog-RSN-1 is now on Amazon SageMaker JumpStart

•8 min read•
#llm#deployment
Level:Advanced
For:AI Engineers
✦TL;DR

PolyAI’s Dialog‑RSN‑1, a real‑time reasoning model for complex spoken dialogs, is now available through AWS Marketplace and Amazon SageMaker JumpStart, enabling one‑click deployment as a SageMaker endpoint. The model consumes 16‑bit PCM mono audio at 16 kHz or 24 kHz, speaks the OpenAI Realtime protocol, and supports connections up to 30 minutes, while remaining English‑only. Engineers can launch the model inside their own VPC with IAM‑controlled access and CloudWatch monitoring, eliminating the need to manage containers or select instance types. This streamlines the integration of live‑call understanding into production dialog agents with full data residency control.

⚡ Key Takeaways

  • Dialog‑RSN‑1 is an English‑only real‑time reasoning model for live calls, accepting 16‑bit PCM mono audio at 16 kHz or 24 kHz and supporting up to 30‑minute connections.
  • Deployment uses the OpenAI Realtime protocol, allowing existing Realtime clients to connect by changing the base URL.
  • One‑click deployment via SageMaker JumpStart automatically provisions a tuned real‑time inference endpoint, removing container handling and instance‑type selection.
  • The endpoint runs inside the customer’s VPC, governed by IAM roles and monitored with Amazon CloudWatch, preserving audio data residency.
  • No manual benchmarking is required; SageMaker JumpStart selects the serving stack for you.
💡 Why It Matters

Engineers can now add high‑accuracy, low‑latency spoken‑dialog capabilities to production systems without building custom inference pipelines, accelerating time‑to‑value while meeting security and compliance constraints.

✅ Practical Steps

  1. Subscribe to Dialog‑RSN‑1 on AWS Marketplace and deploy it directly from the SageMaker JumpStart Model Hub with a single “Deploy” click.
  2. For automated provisioning, use a few lines of the SageMaker Python SDK to launch the real‑time inference endpoint.
  3. Point your existing OpenAI Realtime client to the new endpoint by updating its base URL; feed 16‑bit PCM mono audio at 16 kHz or 24 kHz.

Want the full story? Read the original article.

Read on AWS ML Blog ↗

More like this

Censored vs. Truncated Regression in Python, Explained

Towards Data Science•#llm

New in AWS Quick: More capable flows, an updated web experience, and agents that work together

AWS ML Blog•#agents

Three insights you may have missed from theCUBE’s coverage of the Dell Technologies New York Symposium

SiliconANGLE AI•#deployment

How Google Duplex evolved to also call businesses for you in Gemini

Google Research•#agents

EXPLORE AI NEWS

Daily hand-picked stories on LLMs, RAG, agents and production AI — curated for engineers who ship.

BROWSE NEWS

GET THE WEEKLY DIGEST

Join engineers getting the Monday signal-over-noise AI breakdown. No spam, unsubscribe anytime.

LEARN AI ENGINEERING

Curated courses, research papers, repos and tutorials built for engineers leveling up in AI.

START LEARNING