// yDataLabs · private AI · proven in healthcare

LLMs and AI that run where your data lives.

yDataLabs advises your team and builds the system. AI engineering, platform engineering, and data science from a research-led team.

projects
100+
clients
50+
partners
John Snow Labs, Gesund.ai

// solutions

Three strengths, one team

Most AI projects need all three: someone to build the AI, someone to run it safely, and someone to prove it works.

01

AI engineering

We build AI systems that work in daily use, not only in a demo.

  • LLM applications

    Assistants, search over your own documents, and tools that read and write text for your team. Built on the model that fits the task, open or commercial.

  • LLM integration

    We connect language models to the systems you already run: databases, record systems, and internal tools. Your staff keep their way of working.

  • Agents and workflow automation

    AI agents that carry out multi-step work, such as collecting documents and preparing reports. A person stays in control of every decision that matters.

  • Language and vision models

    Models that pull structured facts out of free text, remove personal details from documents, and read images and scans. Tuned on your own data when needed.

02

Platform engineering

We build and run the platforms that AI needs, inside your own walls.

  • Local LLMs

    A private LLM on your own hardware. We pick the open model, size the servers, and tune speed and cost. No prompt and no document leaves your network.

  • On-premises and air-gapped AI

    Full AI systems on your own servers, including networks with no internet access. Built with access control, audit logs, and the rules of your sector.

  • MLOps and AI platforms

    Platforms to train, test, deploy, and monitor models in a repeatable way. Every model version is tracked, so you can explain and reproduce each result.

  • Data pipelines

    Pipelines that collect, clean, and deliver data to your models and reports. Tested, monitored, and built to handle sensitive records safely.

03

Data science

We turn data into forecasts and decisions that you can explain.

  • Predictive modeling

    Models that forecast outcomes and rank risks from your historical data. For example patient outcomes, demand, or equipment failure.

  • Multi-criteria decision support

    Tools that compare options across many weighted criteria. They make trade-offs visible, so a hard choice about resources can be explained and defended.

  • Multi-agent simulation

    Simulations of complex systems, such as patient flows or supply chains. Test a change in a virtual setting before you make it in the real one.

  • Model evaluation

    Independent testing of AI models before and after launch: accuracy, robustness, bias, and safety. Includes methods to check LLM answers at scale.

// advice

Get advice

Not sure where to start? We help your leaders and your team make good AI decisions before you spend on building.

Book a call
  • AI strategy and roadmap

    Where AI pays off in your organisation, in what order, and what it takes to get there.

  • Feasibility studies

    A short, honest check: can this idea work with your data, your budget, and your rules?

  • Workshops and training

    Hands-on sessions for your engineers, analysts, and managers, built around your own cases.

  • Mentorship and coaching

    A senior expert next to your team during design reviews and hard technical choices.

// proof

Proven in healthcare

Healthcare is the hardest test for private AI. Patient records are among the most sensitive data there is. The rules are strict, and many hospital systems have no internet access at all.

We build and run AI under these conditions: a clinical-grade MLOps platform inside air-gapped environments, de-identification of patient notes, and medical language models at the point of care. What works here works in any industry that must protect its data.

Aug 2021 - Present · Healthcare AI Platform

Clinical-Grade MLOps Platform Development

30% reduction in model deployment time and 100% regulatory compliance for over 50 medical AI models.

2022 - 2023 · Pharmaceutical Company

Advanced NLP for Metastasis Detection

84% F1 score across seven classes, 12% improvement over previous state-of-the-art methods, and 25% reduction in diagnostic time.

2022 - Present · Pharmaceutical Company

Healthcare-Specific LLMs for Oncology Patient Timelines

15% increase in guideline adherence and 40% reduction in manual chart review time.

2023 - Present · Healthcare Provider

Generative AI for Data Extraction Clinical Support

50% faster data extraction process and 20% improvement in data accuracy.

See all work
partnersJohn Snow LabsGesund.ai

// products

Products we built and run

We do not only advise. These are our own products.

Screenshot of ModelPark

ModelPark

Share your AI apps with one link

A platform for data scientists and developers to share and demo their apps. It gives a secure, easy-to-remember address to an app that runs on your laptop or in the cloud.

  • Secure URL Generation with custom, memorable URLs
  • Real-time Monitoring for app status and engagement
  • Advanced Analytics for user behavior insights
  • Privacy Controls with private mode and password protection
Screenshot of MinaMed

MinaMed

Voice-enabled AI medical assistant

An AI assistant for healthcare professionals. It listens, writes the clinical note, and finds the medical reference a clinician needs, so less time goes to paperwork.

  • AI Medical Scribe for automatic transcription
  • Smart Documentation for EHR entry generation
  • Medical Search for knowledge and references
  • Voice Commands for hands-free operation
Dr. Veysel Kocaman

// founder

Dr. Veysel Kocaman

Founder & Director

Dr. Veysel Kocaman is a seasoned Data Scientist and ML Engineer with over a decade of experience leading advanced AI and machine learning initiatives across healthcare, life sciences, and enterprise domains. He is the Director of yDataLabs, a Dutch consultancy that builds AI systems and turns data into business value.

  • Google-recognized ML Developer Expert
  • 40+ peer-reviewed publications in medical AI
  • 100+ talks at conferences worldwide
Read the research

// blog

Latest posts

For leadersPlatform engineering

Local LLM or cloud API: how to choose

Local LLM or cloud API? Start from the data, not from the model. Four questions, with examples from healthcare, help you choose for each task.

For leadersAI engineering

AI in the Clinic, 2026: What Works, What Doesn’t

AI is already in many clinics, but the proof that it helps patients is still thin. What the evidence shows, which tools are in use, and how to adopt AI safely.

All posts

// contact

Book a call

Tell us what you want to build, or where you need advice. A short message is enough.

Location
Maastricht, the Netherlands

Send a message

We use your details only to answer your message. See our privacy page.