Argilla
Argilla builds open-source, human-centric data tools that help AI engineers and domain experts create, curate, evaluate, and maintain high-quality datasets. Its stated mission is to build the open standard for data-centric NLP and make AI more robust, sustainable, accountable, and useful.
At a Glance
- AI/ML engineering and data-science teams
- Enterprises building NLP, LLM, and generative-AI products
- Domain experts and annotation teams
- Healthcare, IT, media, financial services, insurance, aerospace, utilities, and real-estate technology organizations
- +1 more
AI Tools by Argilla
(1)
Distilabel
Synthetic Data Pipeline Framework
Discussions
No discussions yet
Be the first to start a discussion about Argilla
Latest News
Argilla launched no-code creation and curation of fine-tuning and evaluation datasets directly from Hugging Face Hub datasets.
Argilla 2.0 was released with simpler SDK/UI workflows, team collaboration, and easier deployment on Hugging Face Spaces.
Argilla announced the Argilla 2.0 release candidate and its extensible feedback-dataset model.
Argilla announced that it was joining Hugging Face after extensive collaboration on Spaces, open datasets, Notus, ORPO, and community initiatives.
Products & Services
Open-source collaboration platform for AI engineers and domain experts to build and maintain high-quality datasets through human feedback, annotation, evaluation, and data curation.
The 2.0 platform and SDK unify text classification, NER, translation, summarization, QA, LLM, speech, and conversational-AI feedback through configurable fields and questions such as labels, ratings, rankings, free text, and spans.
Open-source Python framework for synthetic-data generation and AI feedback. It lets engineers build serializable, scalable DAG-like pipelines from generators, steps, tasks, and LLM integrations, and share resulting datasets and pipeline configurations.
A collection of fine-tuned language models produced with supervised fine-tuning, DPO, and other RLHF-related techniques, using a data-first approach.
Market Position
Argilla is positioned as an open-source, data-centric collaboration and feedback platform focused particularly on NLP, LLM fine-tuning, evaluation, and RLHF/preference data. Compared with broad annotation platforms such as Label Studio and Prodigy, Argilla emphasizes AI/LLM feedback loops, domain-expert collaboration, programmatic labeling, Hugging Face integration, and data ownership; adjacent competitors include Snorkel for programmatic/weak supervision and Scale AI for managed data operations.
Leadership
Founders
Daniel Vila Suero
Ph.D. in artificial intelligence; co-founder and CEO. He previously worked with the Recognai team on enterprise NLP applications and consulting before focusing full-time on Argilla.
Francisco (Paco) Aranda Montes
Co-founder and CTO. He was part of the Recognai team that built enterprise NLP applications and the Rubrix data-labeling platform before Argilla became the product company.
Amélie Viallet
Co-founder and Head of Product. She helped lead Argilla's product direction, including the transition from Rubrix's task-specific labeling model to Argilla's extensible feedback platform.
Executive Team
Daniel Vila Suero
CEO and co-founder
AI Ph.D.; co-founded Argilla/Recognai and leads its data-centric AI vision.
Francisco (Paco) Aranda Montes
CTO and co-founder
Co-founded Argilla/Recognai and led the technical development of its NLP and data-labeling platform.
Founding Story
Argilla originated as Recognai, an NLP consultancy and product-development company founded in 2017. The team built Rubrix internally to manage data labeling for customer NLP deployments, then open-sourced it in June 2021 after realizing it could fill a gap for practitioners needing a simple, data-centric, human-in-the-loop platform. In October 2022, Recognai and Rubrix were renamed Argilla to reflect the organization's human-centric, collaborative vision.
Business Model
Revenue Model
Open-source software is available free of charge and can be self-hosted or deployed on Hugging Face Spaces. Argilla also developed a hosted/cloud and enterprise offering for organizations needing scaled data-labeling and curation workflows, integrations, and enterprise capabilities.
Pricing Tiers
Apache-2.0 open-source Argilla software; users can run it themselves or deploy it on Hugging Face Spaces.
Target Markets
- AI/ML engineering and data-science teams
- Enterprises building NLP, LLM, and generative-AI products
- Domain experts and annotation teams
- Healthcare, IT, media, financial services, insurance, aerospace, utilities, and real-estate technology organizations
- Open-source AI and Hugging Face community projects
- Language-model fine-tuning and supervised fine-tuning dataset creation
- RLHF, preference optimization, and comparison-data collection
- Model evaluation and continuous monitoring feedback loops
- Text classification, named-entity recognition, translation, summarization, and question answering
- Conversational AI and speech/LLM feedback projects
- Synthetic-data generation and AI-assisted data judging
- Reale Seguros
- Airbus
- Red Eléctrica de España
- Idealista