JetBrains Mellum and the Rise of Specialized LLMs

Livestream
Agenda

As the AI world races toward ever-larger general-purpose models and dreams of AGI (Artificial General Intelligence), a quieter but crucial shift is underway – one that re-embraces the value of specialization, efficiency, and ethical sustainability.

In this livestream, Michelle Frost, AI Advocate at JetBrains, and Vaibhav Srivastav, Head of Community and Platform at Hugging Face, will challenge the dominant “bigger is better” narrative. Together, they explore the case for focal models: small, task-specific LLMs designed with real-world performance, sustainability, and accessibility in mind.

The session features the open-source release of JetBrains Mellum, a focal LLM for code completion, now available on Hugging Face. Learn how focal models like Mellum are reshaping the AI landscape through:

  • Architectural modularity
  • Better performance-to-cost ratios
  • Reduced environmental impact

Whether you’re a developer, researcher, or just AI-curious, join us for a practical discussion on building AI that works – ethically, efficiently, and on purpose.

Speaking to you

Vaibhav Srivastav
Vaibhav Srivastav leads the community and platform efforts at Hugging Face. He studied Computational Linguistics at the Universität Stuttgart. His research is focused on Speech, Language, and Vision methods for extracting insights from unstructured data. Previously, he worked at Deloitte Consulting LLP, advising Fortune 10 tech clients and building scalable cloud solutions as a Google Cloud Platform SME.
Michelle Frost
Michelle Frost is an AI Advocate at JetBrains. With over a decade of engineering experience, Michelle holds a Bachelor of Science in computer science from the University of Missouri and a Master of Science in artificial intelligence from Johns Hopkins University. She is also a Microsoft AI MVP. As an established AI and machine learning specialist, Michelle focuses on responsible AI development. Her approach is grounded in creating AI that is fair, accountable, and transparent.