Data science, AI and training Founded in 2015

Data science & AI
consulting and training.

We help organisations analyse data, develop machine learning and AI applications, and train their teams. Our work draws on statistics, Python, R and open-source technologies, with an emphasis on practical results and knowledge transfer.

stat4decision.py copy
# our main areas of work

from stat4decision import Engagement

engagement = Engagement(
    stack       = ["python", "R", "open LLMs"],  # data science and AI
    packaging   = "uv",            # reproducible environments
    versioning  = "gitlab",        # source control and CI
    deployment  = "on-premise",    # when appropriate
    ownership   = "client",        # agreed project deliverables
    skills_xfer = True             # documentation and training
)

result = engagement.deliver()
# → code, documentation and practical support

Tools and technologies

We work with Python, R and established open-source tools for analysis, modelling, application development and deployment.

🐍Language
Python
pandasscikit-learnPyTorchFastAPILangChainMLflowuvruff
📊Language
R
tidyverseggplot2ShinytidymodelsQuartoforecastrenv
🤖AI / LLM
Open LLMs
Mistral AILlamaHuggingFaceRAGLlamaIndexOllamavLLM
🔧DevOps / MLOps
CI/CD & pipelines
GitLab CI/CDDockerMLflowDVCKubernetesPrefect
📈Dataviz / Apps
Dashboards
StreamlitDashggplot2PlotlyShinyQuarto
🏗️Infrastructure
Posit / R env
RStudio ServerWorkbenchConnectPkg ManagerPositron

Our services

We support organisations through consulting assignments, application development, statistical analysis and professional training.

🤖
Generative AI and LLMs
RAG applications using internal documents, LLM integration, fine-tuning and AI agents. We can work with locally hosted models and take data protection and relevant AI Act requirements into account.
RAGMistral AILangChainvLLMon-premiseAI Act
→
⚙️
Python & MLOps
Development and deployment of machine learning models in Python. We use tools such as uv, GitLab CI/CD and MLflow to manage environments, testing and experiments.
uvGitLab CIFastAPIMLflowDockerDVC
→
📈
Data apps & dashboards
Interactive dashboards and applications with Streamlit, Dash and Shiny. We develop prototypes and applications that can be deployed within your existing environment.
StreamlitDashShinyPlotlyggplot2Docker
→
📊
R & statistical modelling
Statistical analysis and modelling in R, including time series, econometrics and migration from SAS. We also support teams working with Shiny and Posit infrastructure.
RtidymodelsrenvPositSAS migrationQuarto
→
🔬
On-demand data analysis
Targeted analysis for a specific business or research question. Depending on the scope and available data, we can provide results and interpretation within a few working days.
NLPforecastingsegmentationscoringsurvival analysis
→
🎓
Professional training
Training in Python, R, statistics, machine learning, data engineering and generative AI. We offer scheduled courses and tailored sessions for teams. Our training activity is Qualiopi certified.
PythonRMLGen AIQualiopiIn-house and open courses
→

Why we favour open source

Open-source tools offer flexibility, transparency and opportunities to build skills within your own team.

01. Flexibility
More choice over tools and hosting
Python, R and open models allow organisations to adapt their solutions over time. The choice of model, hosting and infrastructure depends on the project.
02. Transparency
Understand and document models
Open-source tools can help with inspection, testing and documentation. Regulatory compliance still depends on how a system is designed and used.
03. Reproducibility
Manage development environments
We use uv for Python and renv for R to record dependencies and make development environments easier to reproduce.
04. Skills transfer
Help teams maintain their work
Documentation, code reviews and training help client teams understand the solutions we develop and maintain them over time.

Clients across different sectors since 2015

We work with private companies, public organisations and research teams in sectors including finance, energy, healthcare and industry.

EDFEngie AXADior L'OréalOrange LCLOCDE UbisoftNestlé ThalesScania CNRSCerba Healthcare CofaceInstitut Curie Société GénéraleIfremer

Discuss your project or training needs

Tell us about your objectives, your data or your training requirements. We aim to respond to enquiries within two working days.

Phone +33 1 72 25 40 82 Email info@stat4decision.com Response within 2 working days