Probabl
Probabl is a Paris-based Inria spin-off that operates scikit-learn, the world's most-used Python ML library (200M+ monthly downloads), and sells Skore, an enterprise collaboration platform for validating, tracking, and governing scikit-learn-based ML workflows, plus training and certification services.
- Company typePrivate
- Founded2023
- HeadquartersParis, France
- Headcount51–100
- GTM typeB2B and B2C
- OfferingSoftware
What Probabl does
Probabl is a Paris-based open-source AI infrastructure company founded in 2023 as a spin-off from Inria, the French national research institute for digital science. The company operates and commercializes scikit-learn, the world's most widely used Python machine learning library, which records over 200 million monthly downloads and is now the default execution layer for agentic AI coding tools. Its product portfolio spans the open-source scikit-learn library, the Skore enterprise collaboration platform (open-source Python library plus cloud-based Skore Hub for experiment tracking, model validation, and governance artifact generation), the Skolar training platform, the official scikit-learn certification program, and professional services including enterprise support contracts.
Probabl's core technology centers on what the company describes as the methodology layer between AI coding tools and production ML systems. Skore validates scikit-learn pipelines for data leakage, picks the right metric per fold, flags silent errors, and auto-generates model cards and documentation. The company has driven adoption of the Python Array API standard in scikit-learn, enabling GPU acceleration through PyTorch, CuPy, and JAX backends with reported 15x speed-ups. Probabl also co-developed the TabICL tabular foundation model with Inria, which delivers a 5x reduction in prediction error versus random forest on the TabArena benchmark. The platform integrates with Jupyter, VS Code, Docker, Kubernetes, and the three major cloud providers.
The business model combines community-led product-led growth with enterprise direct sales. Skore is offered as a freemium self-serve SaaS product (skore.probabl.ai) for individual data scientists and SMBs, alongside behind-firewall enterprise deployments for compliance-sensitive organizations. Additional revenue streams include subscription enterprise support contracts, professional services and consulting (bolstered by the 2024 acqui-hire of Mnemotix), and certification/training fees via the scikit-learn certification program and Skolar platform. Go-to-market leverages the millions-strong scikit-learn developer community through Discord, YouTube (100+ tutorials), the Sample Space podcast, a Substack/Medium blog, and active GitHub engagement. As of January 2026, François Méro (previously of Talend) was appointed CEO, with Yann Lechelle as Executive President/Chairman and scikit-learn co-founder Gaël Varoquaux as Chief Science Officer, formalizing the company's transition into its scale-up phase following a €18.5M total seed financing.
Probabl firmographics
Firmographics- Name
- Probabl
- Legal name
- probabl SAS
- Website
- https://probabl.ai
- Company type
- Private
- Founded year
- 2023
- Operating status
- Operating
- Headcount range
- 51–100 employees
- Short description
- Probabl is a Paris-based Inria spin-off that operates scikit-learn, the world's most-used Python ML library (200M+ monthly downloads), and sells Skore, an enterprise collaboration platform for validating, tracking, and governing scikit-learn-based ML workflows, plus training and certification services.
- Ownership category
- akta.pro rank
Probabl industry classification
Industry- Product category
- Machine Learning Infrastructure
- NAICS
- Software Publishers (513210), Computer Systems Design and Related Services (5415)
- SIC
- Services-Prepackaged Software (7372)
- akta.pro primary industry
- Scientific Data Management & Data Platforms (SDMS/Data Lakehouse for R&D) (HLAGAJAC)
- akta.pro secondary industry
- B2B SaaS & Enterprise Personalization (CRM, Sales, Support) (HDAAAGAJ)
Keywords
Where Probabl is headquartered
LocationHeadquarters
- HQ city
- Paris
- HQ country
- France
- HQ region
- Europe
Offices4 records
Markets served
Probabl business model
Business model- GTM type
- B2B and B2C
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Marketing or Sales, Operations
Revenue model
- Skore SaaS/Premium: Skore platform offered as SaaS for zero-infra teams and behind-firewall deployment for compliance-heavy enterprises. Same product, deployment choice. Likely subscription-based with free tier.
- Training & Certification: Official scikit-learn certification program (Level 2 Professional, Level 3 Expert, Trainer Certification) launched Q4 2024. Skolar platform for structured training courses.
- Professional Services: Enterprise support, expert solutions, and professional services for enterprise AI challenges including acqui-hired talent from Mnemotix.
- Enterprise Support Contracts: Commercial support services for enterprises using scikit-learn and Skore products.
Go-to-market motion3 records
Distribution channels5 records
Marketing channels8 records
Probabl product offering
Product offeringCore offering
Probabl operates and commercializes the scikit-learn ecosystem, the world's most widely used open-source machine learning library with 200M+ monthly downloads. The company sells Skore, an enterprise data science collaboration platform that acts as a methodology layer between AI coding tools and production by tracking experiments, validating scikit-learn models, detecting data leakage, and auto-generating governance artifacts. It also offers Skolar, a training platform for hands-on open-source data science certification, and provides professional services and enterprise support contracts.
Product overview
Probabl operates as a product and services company centered around the scikit-learn ecosystem, offering a portfolio of complementary products. The core offering includes scikit-learn (the foundational open-source ML library with 200+ estimators) and Skore (an enterprise collaboration platform for experiment tracking, model validation, and team workflows). Skore comprises both an open-source Python library and a cloud-based Skore Hub platform. Additional products include Skolar (training platform), scikit-learn Central (community hub), scikit-learn Certification (professional credentials), and Professional Services (training and consulting). The portfolio is designed to support organizations from artisanal data science to industrial-grade AI infrastructure.
Differentiator
Problem solved
Functional benefit
Brands
- Skore: Collaboration platform for data science teams that enables experiment tracking, model validation, and team workflows. Available as open source library and enterprise platform.
- Skolar
- Scikit-learn Central
Products and services
- Skore Enterprise data science collaboration platform that acts as the methodology layer between AI coding tools and production. Tracks experiments, validates scikit-learn models by detecting data leakage, picks the right metric per fold, flags silent errors, auto-generates model cards and documentation. Available as SaaS or behind-firewall deployment for compliance-heavy enterprises.
- skore Python Library Open-source Python library for evaluating and gaining insights from predictive models. Provides EstimatorReport, CrossValidationReport, and ComparisonReport classes with a unified API for model evaluation, inspection, and experiment organization. Distributed via PyPI.
- Skore Hub Cloud-based collaboration platform that extends the skore Python library, enabling teams to track, explore, and share data science workflows with local or remote storage, experiment search, and team visibility features.
- scikit-learn World's most widely used open-source machine learning library with over 200 estimators and 200M+ monthly downloads. Probabl operates and maintains the library as part of its industrial-grade ML infrastructure.
- Skolar Training platform providing hands-on, structured, and certified learning in open-source data science, including video courses and learning paths for scikit-learn proficiency. Offers free and paid courses.
- scikit-learn Certification Program Official certification program for scikit-learn practitioners. Includes Level 2 (Professional) and Level 3 (Expert) certifications, with Trainer Certification for advanced professionals seeking to validate their skills.
- scikit-learn Central Community hub for the scikit-learn ecosystem, providing resources, documentation, and community engagement for the open-source ML library.
- Professional Services Enterprise AI services including training, expert solutions, support, and consulting for organizations adopting scikit-learn and building data science infrastructure.
Quantifiable outcome
- 200M+ monthly downloads of scikit-learn (doubled from 100M in 9 months)
- +3 more outcomes
Companies that use Probabl
Customer profileNamed customers5 records
Segments4 records
Ideal customer profiles4 records
Probabl technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
Integration9 records
AI capability6 records
Feature5 records
Probabl partnerships and signals
Strategic signalPartnerships
Five partnerships are on record, tiered strategic, core and minor.
- IntelstrategicCollaboration announced to accelerate scikit-learn on Intel hardware. Technical partnership for hardware optimization and performance enhancement.
- NVIDIAstrategicJoint technical deep dive on CUDA Python roadmap and Python array API adoption in scikit-learn. Probabl presenting at NVIDIA GTC 2026 conference. Collaboration on GPU acceleration via array API standard.
- QuansightstrategicCollaboration on array API standard adoption in scikit-learn. Quansight contributors working alongside Probabl and NVIDIA on this multi-year effort.
- InriacoreFrench national research institute for digital science. Probabl is a spin-off from Inria; CEO Bruno Sportisse of Inria involved from the beginning supporting the company.
- MnemotixminorAcqui-hired company securing invaluable talent to enhance enterprise client offerings. Talent acquisition to strengthen professional services.
Scale indicators8 records
Recent moves8 records
Expansion highlights6 records
Probabl competitors and assessment
Company assessmentDirect peers
- Neptune.ai: Experiment tracking and metadata store for ML teams, with strong overlap in target persona (data scientists using scikit-learn/PyTorch) and similar product positioning around reproducibility and team collaboration.
- Weights & Biases: Experiment tracking and MLOps platform for ML teams. Directly comparable to Skore as a methodology/observability layer for data scientists, with overlapping enterprise target customers and a similar self-serve to enterprise sales motion.
- MLflow (Databricks): Open-source ML lifecycle management platform including experiment tracking, model registry, and deployment. Listed as a Skore integration, but also a direct substitute for experiment tracking functionality within the Databricks ecosystem.
- Hugging Face: Open-source AI/ML platform company that built a commercial business atop community-led model and dataset sharing. Highly comparable business model: spin-out maintaining a foundational open-source library with enterprise monetization via paid tiers, training, and professional services.
- Anaconda: Commercial Python data science distribution founded by the creator of NumPy (Travis Oliphant, a Probabl investor). Same playbook: monetize the foundational open-source Python data science stack via enterprise subscriptions, training, and services.
- Comet.ml: ML experiment tracking, model comparison, and collaboration platform for data science teams. Direct competitor in the MLOps observability layer, serving similar enterprise data science personas.
- Domino Data Lab: Enterprise MLOps platform with experiment tracking, model governance, and collaboration features. Comparable in targeting large regulated enterprises (financial services, life sciences) that need reproducibility and audit trails.
- DataRobot: Enterprise AI platform with ML governance, model monitoring, and collaboration features. Comparable enterprise go-to-market motion and overlap with Probabl's regulated-industry customer base (banks, insurers).
- Astronomer (Apache Airflow): Commercial company built around Apache Airflow, an open-source data orchestration tool. Comparable open-source commercialization playbook: monetize a community-governed foundational tool via enterprise SaaS, training, and services targeting data/ML teams.
Broad incumbents
- Databricks: Unified data and AI platform with MLflow for experiment tracking, plus full model lifecycle, governance, and deployment. A broad incumbent offering that competes with Skore's experiment tracking as one component of a much larger data/AI suite.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat6 records
Key risks6 records
Key highlights7 records
Customer concentration
Probabl social profiles
Digital presenceProbabl compliance and trust
Trust signalCompliance1 record
Probabl financial estimates
Financial estimateRevenue estimate
Valuation estimate
Probabl leadership team
Management profileNumber of profiles
Profiles6 records
Probabl funding detail
Funding detailFunding overview
Funding rounds3 records
Investors6 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Probabl M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Probabl
What does Probabl do?
Probabl operates and commercializes the scikit-learn ecosystem, the world's most widely used open-source machine learning library with 200M+ monthly downloads. The company sells Skore, an enterprise data science collaboration platform that acts as a methodology layer between AI coding tools and production by tracking experiments, validating scikit-learn models, detecting data leakage, and auto-generating governance artifacts. It also offers Skolar, a training platform for hands-on open-source data science certification, and provides professional services and enterprise support contracts.
Is Probabl a public or private company?
Probabl is a private company. It is classified as venture growth investor backed and is currently operating.
When was Probabl founded?
Probabl was founded in 2023. It employs 51 to 100 people.
Where is Probabl based?
Probabl is headquartered in Paris, France, in the Europe region.
How does Probabl make money?
Four revenue lines are on record. Skore SaaS/Premium is the primary driver. The others are training & Certification, professional Services and enterprise Support Contracts.
Who are Probabl's main competitors?
Direct peers on record are Neptune.ai, Weights & Biases, MLflow (Databricks), Hugging Face, Anaconda, Comet.ml, Domino Data Lab, DataRobot and Astronomer (Apache Airflow). Databricks is listed as a broad incumbent.
Does Probabl have an API?
No public API is recorded for Probabl.
What industry is Probabl in?
Probabl's product category is Machine Learning Infrastructure. Its primary akta.pro industry code is HLAGAJAC, Scientific Data Management & Data Platforms (SDMS/Data Lakehouse for R&D), with a secondary code of HDAAAGAJ, B2B SaaS & Enterprise Personalization (CRM, Sales, Support). Its NAICS code is 513210 and its SIC code is 7372.