Active
Voice Orchestrator
Amazon Connect + Lex + Lambda fulfillment framework
Overview
Context, approach, and outcome.
An opinionated Python framework for building voice-first experiences on Amazon Connect. Handles intent routing, async prefetch, barge-in safety, and context handoff to human agents. Built from scratch after the fifth time I rewrote the same glue code.
Lessons
What I'd carry forward.
Voice budgets are unforgiving. Prefetching on utterance-start beats any downstream optimization. Barge-in changes more than you think.
Stack
Tools and platforms.
PythonAWS LambdaAmazon ConnectAmazon LexDynamoDBTerraform
Related writing
Notes from this project.
- Daily Note: TIL — Polly SSML <mark> tagsPolly's SSML <mark> tags emit timing events over the stream. Useful for synchronizing on-screen captions to voice playback.
- Building Voice Integrations on Top of Async ChatbotsWhat breaks when you front an async chatbot with Amazon Connect + Lex, and how to keep latency, barge-in, and context handoff sane.
Talk shop
Want to discuss this kind of work?
If you're hiring for similar backend, AWS, voice, or integration work — or just want to compare notes on architecture — reach out directly.