DIANNT
WorkAI systemsDharma AI

AI systems / 2025–present

Dharma AI

A conversational study companion grounded in classical Buddhist source material and translated citations.

A multilingual RAG system joining source retrieval, threaded text conversations, and a real-time voice interface.

Dharma AI dashboard with Buddhist text categories and conversation choices
The study dashboard begins with sources, topics, and a choice between voice or text.
Role
AI product and retrieval engineer
Timeframe
2025–present
System
AI systems
  • RAG
  • Python
  • voice AI
  • multilingual retrieval
  • citations
  • knowledge systems
01

Overview

The system in context.

Dharma AI was developed with a nonprofit effort working to preserve ancient manuscripts. The product is designed as a conversational teacher that retrieves classical source passages before composing an answer, keeping study grounded in the underlying texts.

The system brings together multilingual retrieval across Tibetan, Sanskrit, Mongolian, and Classical Chinese material, translated citations, persistent text threads, and speech-to-text and text-to-speech for voice study.

What shipped

  • Grounds generated answers in retrieved passages and presents translated citations.
  • Works across source material in four classical-language traditions described by the project.
  • Supports both threaded text study and a real-time voice interaction path.
  • Extends the annotated-text lineage established during the Asian Classics Institute work.
02

Measured context

The numbers, with their meaning intact.

source-language traditions
4

The portfolio identifies Tibetan, Sanskrit, Mongolian, and Classical Chinese source material.

conversation modes
2

The captured product presents voice and threaded text study experiences.

active product period
2025+

The case study begins in 2025 and is described as ongoing studio work.

03

Product record

Screenshots from the product and project record. Open any frame for a closer view.

01The study dashboard begins with sources, topics, and a choice between voice or text.
02The product surface frames retrieval technology as a calm, source-led learning experience.
04

Build story

From signal to shipped system.

  1. 01 / Preserve

    Start with the source

    Classical material and its translations are organized as the authority layer for every interaction.

  2. 02 / Retrieve

    Ground the conversation

    Questions resolve into relevant passages so answers can carry direct, inspectable source context.

  3. 03 / Converse

    Make study feel immediate

    Persistent text threads and a speech interface turn the retrieval system into an ongoing study companion.

Project links

Open the live product, working demo, publication, or repository.

Related work

All work →