ChaLearn

What ChaLearn does

ChaLearn is a California-based non-profit organization (501(c)(3), operating since 2011) that organizes machine learning challenges and workshops to advance AI research. It does not develop commercial products; instead, it runs competitive challenges across multiple AI domains — automated deep learning (AutoDL/AutoML), reinforcement learning for power grid control (L2RPN), computer vision for human analysis (LAP), meta-learning, connectomics, and medical/HEP applications — to provide standardized benchmarks and accelerate state-of-the-art progress. Challenge hosting is enabled by the open-source Codabench platform (developed with Université Paris-Saclay and maintained by CKCollab LLC) and supporting infrastructure such as the Grid2Op testbed for power grid simulation and the Meta-Album multi-domain benchmark.

The organization is funded by donations, sponsorships, and grants rather than commercial revenue. Sponsors include Google, Microsoft, 4Paradigm, RTE (the French electricity transmission operator), the National Science Foundation (multiple ECCS grants), and the Paris Region (which funded a million-dollar L2RPN 2023 prize). ChaLearn reaches participants through academic conference workshops (NeurIPS, ICML, CVPR, ICCV, AAAI), peer-reviewed publications in IEEE TPAMI and PMLR, dedicated challenge websites, and Google Groups mailing lists. Past challenge websites remain open as ever-going benchmarks, and the AutoDL series alone curated approximately 66 datasets for ongoing benchmark use.

The organization is governed by a board led by President Isabelle Guyon (Google DeepMind), Vice-president Evelyne Viegas (Microsoft), Secretary Anne-Catherine Letournel (Université Paris-Saclay), and Treasurer Kristin Bennett (RPI), with directors spanning Université Paris-Saclay, University of Barcelona, INAOE, RTE, and other institutions. ChaLearn operates as an independent entity with no parent company, no venture capital or private equity backing, and no disclosed commercial revenue.

Name

ChaLearn

Legal name

ChaLearn

Website

https://chalearn.org

Company type

Private

Founded year

2011

Operating status

Operating

Headcount range

1–10 employees

Short description

ChaLearn is a California-based non-profit (501(c)(3)) founded in 2011 that organizes machine learning challenges and benchmarks across AutoML, computer vision, meta-learning, and power grid control, serving ML researchers, students, and industry practitioners through workshops at top AI conferences and the open-source Codabench platform.

Ownership category

akta.pro rank

Product category

Machine Learning Research Challenges

NAICS

Scientific Research and Development Services (5417), Software Publishers (5132)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

AutoML & Low-Code ML Platform Operations (HDAAABAL)

akta.pro secondary industry

Bias, Fairness & Representativeness Testing for Datasets (HDAAALAK)

Keywords

• Machine learning challenges
• Automated deep learning
• Power grid AI
• Computer vision benchmarks
• Meta-learning competitions

Headquarters

HQ city

Berkeley

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2C

Offering type

Services

Cost components

Personnel, Technology or R&D, Operations, Marketing or Sales, Infrastructure

Revenue model

1. Donations and Sponsorship: ChaLearn is a California tax-exempt organization under section 501(c)(3) of the US IRS code. The organization raises funds through donations and sponsorships to organize new events and offer prizes and travel awards to deserving participants. Past and present sponsors include NSF, Google, Microsoft, 4Paradigm, Orange Labs, IEEE Computational Intelligence Society, and others.

Go-to-market motion1 record

Distribution channels2 records

Marketing channels4 records

Core offering

ChaLearn organizes and hosts machine learning research competitions across domains including automated deep learning (AutoML/AutoDL), power grid control (L2RPN), computer vision for human analysis (LAP), meta-learning, connectomics, medical applications (MediChal), and physics fairness (Fair Universe). Challenges are run on the open-source Codabench competition platform with automated code submission and evaluation against concealed datasets, and past challenge websites remain open as evergreen benchmarks.

Product overview

ChaLearn is a non-profit organization (501(c)(3)) that organizes machine learning competitions and workshops to advance AI research. The core offering is a challenge hosting platform built on Codabench, which provides automated code evaluation on concealed datasets. ChaLearn's main challenge series include: (1) AutoDL for automated deep learning across multiple modalities (images, video, text, speech), (2) L2RPN for AI control of power grids, (3) LAP (Looking at People) for computer vision human analysis, (4) Meta-Learning challenges for few-shot learning, and (5) domain-specific challenges like Connectomics and MediChal. Supporting products include Grid2Op testbed platform, Meta-Album benchmark datasets, AutoDL Starting Kit with Jupyter notebooks, and the CiML (Challenges in Machine Learning) book series published by Springer. The organization has been running competitions since 2011, partnering with major conferences like NeurIPS, CVPR, and ICCV.

Differentiator

Problem solved

Functional benefit

Products and services

• ChaLearn Challenge Platform Cloud-based platform for organizing and hosting machine learning competitions, enabling participants to submit code that is automatically trained and evaluated on concealed datasets. Built on the open-source Codabench competition platform and serves as the technical backbone for all ChaLearn challenge series.
• AutoML/AutoDL Challenge Series Series of automated deep learning competitions covering image recognition (AutoCV), speech recognition (AutoSpeech), natural language processing (AutoNLP), time series (AutoSeries), and multi-domain challenges with any-time learning metrics. Targeted at ML researchers developing automated deep learning pipelines.
• L2RPN (Learning to Run a Power Network) Challenge series on controlling the French electricity transmission grid using reinforcement learning and AI, addressing real-world problems of carbon neutrality and renewable energy integration. Targets reinforcement learning and operations research researchers.
• LAP (Looking at People) Challenge series pushing state-of-the-art in computer vision for detecting, recognizing, and interacting with humans, including gesture recognition, face anti-spoofing, and body pose estimation. Targeted at computer vision researchers.
• Meta-Learning Challenges Series of few-shot learning competitions including MetaDL at NeurIPS/AAAI, cross-domain meta-learning, and meta-learning from learning curves for rapid model adaptation. Targeted at meta-learning researchers.
• Codabench Competition Platform Open-source competition platform developed and actively supported by ChaLearn for hosting benchmark challenges with automated code submission, evaluation, and leaderboards.
• ChaLearn Connectomics Challenge Neural connectomics challenge on recovering neural network structure from patterns of neural activity (calcium imaging data), addressing brain wiring reconstruction. Attracted 143 participating teams.
• Fair Universe Challenge Challenge series on bias and systematic errors in High Energy Physics simulations, addressing fairness in scientific computing. Targeted at ML researchers working on fair and robust AI for science.
• MediChal Series of medical application challenges including EEG analysis competitions for healthcare machine learning. Targeted at researchers applying ML to medical data.
• CiML Book Series Challenges in Machine Learning book series published by Springer covering AutoML, gesture recognition, connectomics, and competition methodology, including free downloadable PDFs. Edited by ChaLearn organizers.

Quantifiable outcome

• Challenges organized across multiple domains including AutoML, AutoDL, L2RPN (power grids), Connectomics, Looking at People, Meta-learning, and Causality
• +1 more outcomes

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability14 records

Feature2 records

Partnerships

Six partnerships are on record, tiered core and minor.

• GooglecoreGTM or Marketing PartnerGoogle is the primary sponsor of the AutoDL challenge series, helping define tasks, protocol, and data formats. Google also provides cloud resources and co-organizes challenges with ChaLearn.
• 4ParadigmcoreGTM or Marketing Partner4Paradigm donated prizes, datasets, and contributed to the protocol, baseline methods and beta-testing for AutoDL challenges. They co-organized KDD Cup 2019 on temporal relational data.
• RTE (Réseau de Transport d'Électricité)coreStrategic or Co-development PartnerRTE, the French electricity transmission system operator, partners with ChaLearn for the L2RPN (Learning to Run a Power Network) challenge series on real-world power grid control problems. L2RPN 2023 had a million dollar prize from the Paris Region.
• Université Paris-SaclaycoreStrategic or Co-development PartnerUniversité Paris-Saclay administers the Codalab competition platform used by ChaLearn. Many ChaLearn organizers and contributors are affiliated with the university.
• CKCollab LLCcoreTechnology or IntegrationCKCollab LLC maintains the Codalab competition platform with primary developers Eric Carmichael and Tyler Thomas.
• SpringerminorStrategic or Co-development PartnerSpringer publishes the Challenges in Machine Learning (CiML) book series edited by ChaLearn organizers, documenting challenge designs and results.

Scale indicators5 records

Recent moves6 records

Expansion highlights6 records

Market position

Competitive moat4 records

Key risks5 records

Key highlights6 records

Customer concentration

Compliance1 record

Revenue estimate

Valuation estimate

Number of profiles

Profiles15 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about ChaLearn