Hugging Face just dropped a fresh batch of internship openings, and the framing is notably different from your standard tech job post. The company is explicitly prioritizing applicants from marginalized communities — people of color, working-class backgrounds, women, and LGBTQ+ folks — arguing that the people most likely to be hurt by AI’s failures should be the ones shaping its development. Seven positions are up for grabs across three teams.
The Open Source team is hiring for two roles: an Accelerate intern to lead feature integration and a Text to Speech intern focused on reproduction work. The Science team has three slots — Embodied AI with reinforcement learning in simulators, a Fast Distributed Training Framework for large language models, and a Datasets for LLMs position aimed at building the next generation of training data. Rounding out the list are two cross-disciplinary roles: Social Impact Evaluation and AI Art Tooling.
What stands out is the deliberate structure. Each intern works directly alongside library maintainers or research leads, not in some isolated side project. The Social Impact Evaluation role is particularly interesting — developing a technical framework to assess generative models’ societal effects is the kind of work that’s been talked about for years but rarely operationalized. Hugging Face is essentially saying: we’ll pay someone to figure out how to measure harm.
Locations vary by position, and the application process requires a Hugging Face account with a short submission at the end of each form. The company’s equal opportunity language is standard fare, but the upfront mission statement — democratizing responsible machine learning — signals this isn’t just about filling seats. It’s about who gets to build the tools that everyone else will use.
