Serverless Audio Transcriber with Rust and Python Using Pulumi | glowing telegram - Episode 134
In this video, I walk through the process of making our Rust and TypeScript web app, Glowing Telegram, serverless using Pulumi and Python. We’re specifically focusing on developing an audio transcriber that utilizes AWS services. This involves experimenting with different instance sizes in AWS to optimize performance for speech-to-text tasks.
First, I dive into the development of our batch job setup, which aims to seamlessly transcribe audio files. By leveraging AWS configurations, speech language settings, and an S3 client, we ensure that our application can handle various input options. I also discuss the use of the Whisper speech recognition system, exploring different models and the potential for optimizing our processes.
Throughout the video, I address the practical implementation of handling output and data integration with DynamoDB. This includes how our program writes transcribed data into the database while considering considerations for data consistency and order of execution.
Additionally, I delve into the technicalities of deploying this setup with Pulumi. We discuss Docker image usage, the configuration of Pulumi components, and how they interact with AWS services, providing insights into potential abstraction for reusable infrastructure.
Tune in to learn more about integrating Rust with serverless technologies using Pulumi!
🔗 Check out my Twitch channel for more streams: https://www.twitch.tv/saebyn
GitHub: https://github.com/saebyn/glowing-telegram
Discord: https://discord.gg/N7xfy7PyHs