Machine Learning & AI Consultant
Sopra Steria

At the moment, I help organisations work out where machine learning and AI are genuinely useful. Most of my time currently goes into geospatial computer vision.
At the moment, I spend most of my time on computer vision. I’ve also worked with language models, data platforms, and the software that keeps it all running.

Before I choose a tool, I want to know what actually needs to change.
I ask the people involved what they want to achieve and what gets in the way. A good “why?” usually gets us to the real problem.
Then I look at the process, existing software, available data, and technical constraints. Once I know what the solution needs to work with, I start building.
I started in software engineering, went to university for computing science, and gradually moved into applied AI. Here is the short version, from now back to 2012.
Read my 2021 paper on robust image classification ↗Experience from 2025 back to 2012
Sopra Steria

At the moment, I help organisations work out where machine learning and AI are genuinely useful. Most of my time currently goes into geospatial computer vision.
Valtes

I worked on AI features for an app that supports informal caregivers. That meant going from early experiments to code that could live inside a real product.
A few things I worked on: AI features for caregivers · Process automation · Product engineering
Researchable

I helped clients work out where LLMs and RAG could help, then built the applications around them. I also maintained public-data platforms used by the Urban Mobility Observatory.
A few things I worked on: Client workshops · LLM and RAG applications · Public-data platforms
University of Groningen

I specialised in Intelligent Systems and Visual Computing. My thesis explored whether COSFIRE features could improve CNN image classification.
A few things I worked on: Intelligent Systems · CNN classification · Published research
Wild Sea Studio

I built custom Laravel applications, helped colleagues when they got stuck, and kept the cloud-hosted products running.
A few things I worked on: Custom applications · Team mentoring · Cloud infrastructure
Parkos

I built web-scraping infrastructure with headless browser clusters and Selenium Grid. It collected pricing data from more than forty nodes and fed the internal dashboards.
A few things I worked on: Browser clusters · Pricing data systems · Operational dashboards
BeepRoger

I built maritime web applications and followed them from a designer’s first version through to delivery for the customer.
A few things I worked on: Web applications · Product delivery · Customer collaboration
University of Groningen

I took a bridging programme from applied software engineering into academic computing science. In practice: plenty of mathematics, algorithms, and programming.
A few things I worked on: Academic foundation · Mathematics · Computer science
Hanze University of Applied Sciences

This is where I learned the software engineering fundamentals: programming, architecture, project work, and building things with a team.
A few things I worked on: Software fundamentals · Object-oriented design · Project work
I built one with Gradio and OpenAI. It uses my work history, education, and technical experience to answer questions about what I’ve done.
“What has Joey built with computer vision?”
I’ve played classical piano for more than fifteen years. There is always another passage to work on and another detail to fix. Software is not exactly the same, but close enough.
See what I’m practising on Instagram
Have a role, project, or slightly complicated problem involving AI, data, or software? Send me a message and tell me what you’re working on.
j.p.antonisse@live.nl