Skip to Main Content
Lightning Talk Intermediate First Talk

Crafting Tamil Datasets: A Journey from Text to Voice

Proposal status is Rejected
Session Description

In this session, I will take you through my journey of creating structured datasets for Tamil literature, including the revered Thirukkural and Aathichoodi. I will share how I transformed unstructured Tamil texts into organized datasets, enriched them with voice data, and made them accessible for AI applications. I will also talk about the challenges I faced in preserving linguistic nuances and how I tackled them using Python and other tools.

Whether you are a data enthusiast, a developer, or someone passionate about Tamil, this session will provide practical insights and step-by-step guidance on building and sharing datasets that celebrate Tamil literature.

Key Takeaways:

  • Extracting and structuring Tamil texts effectively.
  • Adding voice support to Tamil datasets using Python.
  • Challenges in maintaining linguistic authenticity and solutions.
  • Tips on making datasets open-source and encouraging community contributions.


References

https://www.linkedin.com/posts/selvakumarduraipandian_thirukural-ai-nlp-activity-7297457568971464704-84es

Session Categories

FOSS

Speakers

Selvakumar Duraipandian
Selvakumar Duraipandian Senior Solution consultant | Kissflow

I am a Senior Solution Consultant at Kissflow and one of the developers of Thirukural.ai. With a passion for preserving and promoting Tamil literature, I have created multiple datasets, transforming classical texts into structured formats enriched with voice support. My expertise in Python and data engineering has enabled me to make Tamil datasets accessible for AI and machine learning applications. I would encourage community contributions to expand the reach of Tamil literature in the digital era.