Iztok Bajec is an Associate Professor at the University of Ljubljana with eight years of professional experience intersecting academia and machine learning engineering. He contributes to high-profile open-source projects like NVIDIA NeMo, where he has improved large-language-model and speech AI training workflows, added Camembert model support, and fixed multi-node SLURM resume issues—practical work that bridges research and production-scale systems. Based in Slovenia, he blends teaching and research with hands-on debugging and data-processing improvements for tarred datasets and distributed GPU training. Colleagues value his ability to translate complex academic ideas into robust, usable tooling that advances both student learning and real-world ML deployments.
A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
Role in this project:
ML Engineer
Contributions:30 reviews, 5 commits, 15 PRs in 6 months
Contributions summary:Iztok primarily contributes to the NVIDIA/NeMo project by addressing issues related to training and data processing for large language models and speech AI. They fixed a bug in the training resume functionality within a SLURM multi-node multi-GPU environment, ensuring proper file management across processes. Furthermore, the user added support for the Camembert Huggingface bert-like models to the framework, improving the flexibility for different models. Additionally, they updated documentation and fixed issues within the tarred dataset processing.
Contributions:50 pushes, 14 branches in 2 years 6 months
nlpconversationalconversational-aichatbotnemo
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.