Jointhubs

AI/ML laboratory

We help you
enjoy your data
We help you
enjoy your data

We build AI systems that help people do their work.

ALL YOUR ANSWERS ARE HERE. WAITING.

SCROLL

How we can help

Three ways to work together

Start with what hurts today, and we will shape the rest around your budget and pace.

Process automationRepeatable tasks, documents and reports flow without retyping them by hand.See the offer →

// PROJECTS

Portfolio

Product

Finance

Mamut

Daily research on Warsaw Stock Exchange companies in one panel.

  • News from many sources, full text and sentiment
  • Top stories: only importance 6 and above
  • Topic map, price charts and indicators

StatusRunning · we use it daily

Talk to us about it
Product

Healthcare

Neurohubs

Automated reading of qEEG, built with a neurotech team.

  • Automated qEEG reading
  • Comparison with normative profiles
  • Rule-based engine, AI on the roadmap

StatusPrototype · demo online

Open Neurohubs
Product

AI platform

Neo

An AI agent workspace on your own machine.

  • Runs fully locally, on your machine
  • Watches competitors, registers, documents and receivables
  • Built-in sign-in and encryption

StatusDeployed at a client · running

Book a demo
Prototype

HR and payroll

Pay transparency

Job evaluation and the gender pay gap report in one app.

  • AI proposals backed by quotes from job descriptions
  • Facilitator panel: department progress live
  • Gender pay gap, 5% threshold

StatusIn build · pilot with a client

Talk to us about it
Prototype

Time series

Intelligent forecasting

Forecasts of prices, sales and demand that see the context.

  • Time series: prices, sales, demand
  • One model of the whole, nested by category
  • A separate model for each segment

StatusProven method

Talk to us about it
Prototype

Integrations

Workflow automation

AI wired into the tools your team already uses.

  • Workflows in the tools you already use
  • Semantic search over documents
  • Reusable integration blocks

StatusProven method

Talk to us about it
Open

Squads of AI agents

Fenix

Squads of AI agents for software development.

  • Planning, code, review and tests run by agents
  • Tasks and decisions in Linear
  • Standalone and supervised modes

StatusOpen source · we use it daily

View on GitHub
Open

Knowledge harness

Jointhubs OS

Your Obsidian second brain, open to your agents too.

  • Obsidian, locally on your machine
  • MCP server in Docker over the same vault
  • Works with any agent system, including Claude Code

StatusOpen source · we use it daily

View on GitHub
Open

Data science · Library

Statistical Context Engineering

Context-aware statistical signals for models.

  • Context-aware features
  • Leakage-proof validation
  • Hierarchical statistical signals

StatusOpen source · PyPI

View the library
Open

Data science · Paper

ROE Framework

Multi-criteria validation for forecasting models, published in Risk.net.

  • Three metrics: WSA, WAPM and CESDM
  • Correlation-adjusted variants
  • Case study: GDP forecast in three scenarios

StatusPaper

Read the paper
1 / 1
Mateusz Stachowicz

Mateusz Stachowicz

Founder

View on LinkedIn

For over 9 years I have worked with artificial intelligence systems: I designed, built and maintained them for the largest corporations in the world. I turned that experience into my own methodology, proven in practice and working every time.

  1. Context

    Where and in what environment the solution has to work.

  2. Data

    What the data is and what it really says.

  3. Method

    Choosing the approach and the smallest useful way forward.

// COLLABORATION

What others say

From people we have built artificial intelligence systems with.

“Mateusz showed a high level of expertise in designing AI systems, from hardware architecture through model selection to implementation. He could explain complex technical issues precisely and clearly to people without specialist IT knowledge.”

Dr Krzysztof R. Kupiński

Rector

WSKM in Konin

LinkedIn

“Together we built Neurava, which won NeuroHack Wrocław 2024. We shaped the product concept as a team, and Mateusz designed the AI and data-processing architecture that made the neurobiological signals usable.”

Maurits Unkel

Postdoctoral Researcher

Erasmus MC Rotterdam

LinkedIn

“Mateusz turned our requirements into a clear app at lightning speed, one that makes job evaluation much easier.”

Adriana Jagiełka

Deputy Head of HR and Payroll

Vinci & Vinci

LinkedIn

“Mateusz turned legal requirements from regulations and internal documentation into concrete modules of a working application that meets the strictest standards.”

Kamil Berych

Vice President

Altiorem Foundation

LinkedIn

Collaboration

Engagement Tiers

From a single conversation to full production.

The first 20 minutes of the call are free. Book 20 minutes → (opens the booking window)

Introduction

€50 / h
After the first 20 free minutes

Let’s chat.

Start here

Tell me about your situation:

  • what you have in mind
  • what bothers you
  • what you need

We’ll take it from there.

Production

It depends
Every consultant says that. After 20 minutes I'll tell you on what.

Deploy, scale, maintain.

Production preview
Discuss scope
  • Enterprise SLA
  • System integration
  • Compliance & security

BEFORE YOU ASK

The questions you are actually asking.

What does it cost to find out if this is worth doing?

The first 20 minutes of the call are free. After that you can order a Discovery hour to name the real question and judge the feasibility before anything is built. If it turns out you do not need me, you will know that quickly and cheaply.

You work across property, contracts, clinical data and delivery. Does that not make you a generalist?

The domains differ. The method does not. Each one runs the same steps in the same order, and the domain knowledge comes from your team. What I bring is the part that is identical every time, which is exactly the part most projects get wrong.

Who owns the code when we stop working together?

You do, from day one. Your cloud project, your database, your repository. Nothing runs on my login.

Our data cannot leave the company.

Then it does not. Local-first is the default, not the upsell. Neo runs entirely in Docker on your own machine, and Jointhubs OS embeds locally with cloud calls only when you switch them on.

Share your vision. Let's build it together.

One message. I will tell you honestly whether this is a real problem for AI, a plain software problem, or something you should not build at all.