Streamlining AWS Lambda with Rust and Python | glowing telegram - Episode 140
In this video, I delve into the intricacies of enhancing the Glowing Telegram project by configuring AWS Lambda functions using Rust and Python. We begin by addressing some issues encountered in previous streams, particularly focusing on the setup involving GPU resources and the challenges faced with DynamoDB’s storage constraints when handling audio transcriptions via OpenAI's Whisper.
We'll explore the concept of processing video files with step functions to orchestrate the transcription and summarization pipeline. This involves leveraging OpenAI's API for creating structured outputs and implementing JSON schema to enable seamless, schema-compliant data retrieval. I discuss the development and refinement of prompt engineering strategies to optimize these processes.
Additionally, I tackle challenges related to deploying AWS Lambda functions using Pulumi, transitioning from interactive to batch processing for more efficient workflows, and coding a Lambda function from scratch.
🔗 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