ML Research Engineer Internship, OS Agents - EMEA Remote

Hugging Face, a leading AI platform company with over 5 million users and 100k organizations, is seeking an ML Research Engineer Intern for their innovative OS Agents project. This role focuses on the cutting-edge development of LLM agents - systems that connect Large Language Models to the real world through actions and observations.

The internship specifically centers on developing agents that can interpret screen recordings from any GUI and interact through mouse/keyboard interfaces. This groundbreaking project sits at the intersection of post-training, inference, orchestration, and UX design, offering a unique opportunity to work with state-of-the-art technology.

As an intern, you'll be integral to improving the agent framework, preparing fine-tuning datasets, developing training recipes, and managing inference orchestration. You'll have access to advanced computing resources, including CPU and H100 clusters for large-scale processing and model training. The role emphasizes open-source development, with all code and models being released to benefit the wider AI community.

Hugging Face offers a supportive, diverse, and inclusive work environment where continuous learning is encouraged. The company provides comprehensive benefits including flexible working hours, remote work options, and educational support. You'll work alongside industry experts in a company that values scientific advancement through collaboration.

This is an exceptional opportunity for someone passionate about open-source development, making complex technology accessible, and contributing to one of the fastest-growing ML ecosystems. While specific technical requirements are flexible, enthusiasm for AI and a creative approach to problem-solving are essential.

The role is part of Hugging Face's ambitious 2025 initiative focusing on agent development, positioning you at the forefront of AI innovation. You'll be contributing to a company whose open-source libraries have garnered over 400k+ stars on Github and whose platform hosts over 1M models, 300k datasets, and 300k apps.