BHASHINI Launches Hackathon to Drive AI Model Training for Rajasthan’s Regional Dialects

New Delhi, Jul 3 (KNN) The Digital India BHASHINI Division (DIBD) launched the Rajasthan Language Model Training Hackathon to strengthen AI models for regional languages and expand the use of language technology in digital governance, while also showcasing its broader multilingual artificial intelligence (AI) solutions at the National Conference on e-Governance (NCeG) 2026.

The conference featured live demonstrations of BHASHINI's AI-powered speech and text translation, voice-based services, ShrutLekh, BhashaDaan and Suno Sutra, India's first handheld multilingual edge AI device. 

Delegates also experienced governance applications including CPGRAMS, Nyaya Setu, Arogya Vani, SabhaSaar and Yuva Saarthi, demonstrating how multilingual AI is improving access to public services across languages.

The Rajasthan Language Model Training Hackathon was launched to create high-quality language datasets for Marwari, Mewari, Dhundhari, Hadoti, Mewati and Bagri, with the objective of strengthening AI models for Rajasthan's regional languages.

BHASHINI also deployed ShrutLekh, its AI-powered real-time speech-to-text solution, across plenary and breakout sessions at the conference, enabling live transcription and translation to improve accessibility during technical discussions.

Visitors at the exhibition participated in BhashaDaan, BHASHINI's national crowdsourcing initiative, by contributing and validating language datasets across Indian languages. Participants received digital certificates in recognition of their contributions to India's multilingual AI ecosystem.

The exhibition was visited by Rajasthan Chief Minister Bhajan Lal Sharma, Information Technology and Communication Minister Rajyavardhan Rathore, senior government officials and delegates, who interacted with BHASHINI's multilingual AI solutions.

On the sidelines of the conference, BHASHINI Chief Executive Officer Amitabh Nag participated in the Investors' Meet and held discussions with the Rajasthan Chief Minister on advancing multilingual AI and language technologies to support inclusive digital governance in the state.

Nag said AI must understand the languages people speak in their daily lives to truly serve every citizen, adding that language data contributions, validated datasets and new use cases are helping build more inclusive and accessible AI systems.

According to the Ministry of Electronics and Information Technology (MeitY), BHASHINI continues to expand India's multilingual digital public infrastructure by enabling governments and institutions to deliver services in citizens' preferred languages. 

The platform currently supports over 800 government websites, processes more than 20 million AI inferences daily, has recorded over 8 billion cumulative AI inferences, and offers services across 36 Indian text languages, 23 Indian voice languages, 35 international languages, along with more than 20 specialised natural language processing (NLP) services.

(KNN Bureau)