almoliky.dev
All work

Selected work

06 / 09

Let’s Summarize

An AI summarization platform built in a seven-day hackathon sprint under the letssummarize organisation — YouTube videos, documents and raw text in, summaries out.

The Let’s Summarize home screen: a field for a YouTube URL above options to summarize documents or text.
The home screen: paste a YouTube link, or switch to documents or text.

I designed the core backend architecture: swappable models (GPT-4o, DeepSeek-V3, Gemini), Whisper transcription pipelines, text-to-speech, S3 storage, and a public developer API where integrators bring their own keys.

Summaries can be customized by length, format, model and language, in a fast or a slow mode, and the platform takes PDF, DOCX and TXT files as well as video.

What I built

Backend

  • Designed and implemented the core backend architecture for the platform.
  • Implemented support for multiple language models, including OpenAI, DeepSeek and Gemini.
  • Built video transcription and audio processing pipelines.
  • Implemented document and file upload handling for summarization.
  • Integrated text-to-speech generation and Amazon S3 storage.
  • Designed and exposed a public developer API with API key support.