AI Research Intern

About this internship

Skills you'll use

What you'll do

• Contribute to pretraining, fine-tuning or alignment of Indic language models
• Curate, clean and evaluate large multilingual training datasets
• Design and run evaluation benchmarks for Indian language tasks
• Work on speech recognition, text-to-speech or speech translation models
• Optimise inference for cost, latency and on-device deployment
• Run and analyse distributed training experiments on GPU clusters
• Document research findings and share results with the team

Who can apply

• BTech/MTech/MS/PhD in CS, EE, Mathematics or Computational Linguistics
• Strong foundations in machine learning, deep learning and linear algebra
• Hands-on experience training models in PyTorch
• Understanding of transformer architectures and attention mechanisms
• Experience with NLP, speech processing or multimodal learning
• Comfort working with large datasets and data quality problems
• Familiarity with multi-GPU or distributed training is an advantage
• Demonstrable projects, publications or open-source contributions
• Interest in Indian languages and their computational challenges
• Available full-time on-site in Bengaluru

Perks & benefits

Interview & selection process

4 rounds — typical selection flow for this role:

1. 1Application review including projects, publications and code portfolio
2. 2Technical screen — machine learning fundamentals and coding
3. 3Deep-dive interview on deep learning, transformers and research reasoning
4. 4Final discussion with the research team on problem approach and fit

About Sarvam AI

Searching for “Sarvam AI internships” or “internship at Sarvam AI”? You’re in the right place.