My contribution

Designed and built a voice-based book knowledge assistant using OpenAI file-based retrieval, speech-to-text, and text-to-speech. Custom retrieval and agent-based capabilities are future roadmap work.

Technical approach

Built a voice-based book assistant using OpenAI file retrieval and speech-to-text / text-to-speech. A custom RAG pipeline and agent-based features are roadmap items.

  • Use OpenAI file-based retrieval to ground answers in uploaded book content.
  • Connect speech-to-text and text-to-speech around the question-and-answer workflow.

Simplified voice and retrieval workflow

  1. 01Spoken question
  2. 02Speech-to-text
  3. 03OpenAI file retrieval
  4. 04Text-to-speech answer
DharmaAI: simplified illustration of the documented workflow or implementation areas. Infrastructure and confidential implementation details are intentionally omitted.

Outcome & lessons

Grounded answers use uploaded book content through OpenAI retrieval.

Custom RAG infrastructure and agent-based features are roadmap work. The current description covers file retrieval and voice interaction.

Project context & evidence

Independent application documented in the supplied resume. No production usage figures are claimed.

A public repository or demo link has not been provided for this project.

← More selected workDiscuss a similar challenge ↗