Summer LLM Internship Opportunity with Swecha for Telugu Language Enthusiasts!

Swecha, a nonprofit organization dedicated to the advancement of Free Software and Free Knowledge, is offering an incredible summer internship program focused on developing Telugu Language-centric Large Language Models (LLMs). This unique initiative, in collaboration with IIIT Hyderabad, Ozonetel, and TASK, aims to blend artificial intelligence (AI) with local Telugu culture.

Internship Highlights:

• Project Focus: Developing Telugu Language Models with an emphasis on local songs, food, folk tales, and traditional skills.
• Learning Objectives:
  • Participants will receive hands-on training with Llama from Meta (Facebook).
  • Interns will work on generating LLMs that can process and understand Telugu text with higher speed and accuracy.
  • The project includes building a Text-to-Speech (TTS) model, creating your own Voice Avatar.
  • The scope extends to collecting a Telugu corpus for future language tech.
  • Contribute to building the world’s first supercomputing cluster for AI.
• Skills Developed: AI fundamentals, data collection, language processing, TTS modeling, and cultural integration.

Program Structure:

• Mode: Online
• Time Commitment: Approximately 60-90 minutes per day.
• Training: The first 3 weeks will be dedicated to training.
• Project: The last week will be focused on a project, where participants will work on building a Telugu software based on LLMs.

Eligibility: This internship is open to students, professionals, and anyone interested in AI, language processing, and in Telugu culture.

Application Details:

• Start Date: Tentatively from 24th May 2024 for batch 2.
• Duration: 1 month (3 weeks training + 1 week project)
• Mode: Online
• Fee: Rs 199 (a nominal fee)
• How to Apply: Interested candidates can apply online at https://swecha.org/summer-of-ai

Additional Information: Participants will be part of the largest crowdsourced AI effort for preserving Telugu culture, providing a unique learning experience that enhances both AI skills and cultural appreciation. Top performers will be awarded Swecha fellowships.

Inputs from 1516

Large Language Models work on the basis of existing texts that are written. The machine basically reads how one word follows the next and learns how humans speak language and write language. By building a larger corpus of texts, we can help create a ChatGPT-like model for Telugus which is more focused on how Telugu works and is suitable for 10 crore Telugu people in twin Telugu states and abroad.

It would also eventually help in aiding the learning of Telugu for our future generations by helping them use voice chat with Telugu LLMs.

This isn’t an exact internship per se by the looks of it (they are asking for a nominal fee of Rs 199 instead of paying us). I believe this is to cover the basic costs of some experts and to weed out people who are not serious. They could also be focused on crowdsourcing everything which means they need more people to get better results.

Should you opt for this?

You must understand that this sort of training in “How to train LLMs” is priceless. LLMs are all the rage right now and and knowledge of how to train LLMs is definitely an advantage in your career. It also adds something to your resume that others lack. Llama is an open-source LLM released by Facebook, and knowledge of Llama opens up more opportunities for you in the AI space.

Personally, as a advocate to further the cause of Telugu, I would recommend that you enroll in this. It would benefit everyone in the process and I don’t think 200 bucks is a deal breaker for anyone.

Contact: For any questions, call 04045210808. It will lead to an IVR and you have to wait for 2-3 minutes to get to a human volunteer that can answer your questions. For email, you can reach out via [email protected].

Don’t miss this opportunity to be at the forefront of AI and language technology! Apply now and contribute to the future of Telugu language computing.

You can sign up here at the Swecha Summer of AI.