Boardroomto production.
I’m Gaétan Vandenberghe. I find where AI can genuinely improve a process, build the system behind it, and make it dependable with validation, evaluation and human review.
01Selected work
Systems, not demos.
-
01Co-libry · 2024
Real-estate publishing engine
Property photos, source data and domain knowledge turned into structured, validated listings for CMS and ERP publishing.
Read case study -
02Nordic Collection · 2025–now
AI company platform
Internal tools connecting operational data, APIs and AI-assisted workflows, with business rules and human review.
Read case study -
03Engineering approach
Reliable LLM workflows
Validation, evaluation, bounded retries and human review around model output.
Read case study -
04Model adaptation
Domain-specific fine-tuning
Dataset preparation, baseline comparison and failure analysis for specialised language models.
Read case study -
05Computer vision & robotics
Crop-monitoring drone
Autonomous flight, image capture and object detection for crop monitoring.
Read case study
02Principle
Reliable AI comes from the system around the model.
03Approach
Business context. Engineering detail.
I work between the business question and the technical answer: where AI can genuinely improve a process, the system behind it, and the people who need to trust it.
- 1
Map the process first
Who performs each step, which systems hold the data, where decisions are made and where errors enter. That map becomes the technical design.
- 2
Constrain the output
Ask the model for explicit structured output, and validate it against a schema before downstream code accepts the result.
- 3
Evaluate, don’t guess
Representative evaluation cases expose recurring failure modes and make prompt, model or pipeline changes comparable.
- 4
Keep people in the loop
Repair what’s clear, retry within bounds, and escalate ambiguous or consequential cases to a person with the context.
04Experience
Advising and building, usually both.
Jul 2026 – Present
Founder, AI Advisor & Engineer
Algeevra (opens in a new tab), Belgium
Helping organisations build AI confidence: employee education, opportunity discovery and roadmaps people actually adopt.
Mar 2025 – Present
AI Advisor & Engineer
Nordic Collection, Ghent
Python automations, LLM-assisted workflows and full-stack internal tools (React, TypeScript, Node, PostgreSQL/Supabase), plus AI strategy advice for management and VC investors.
Feb 2024 – Aug 2024
AI Engineering Intern
Co-libry, Ghent
Helped build an early LLM pipeline that turned property photos into CMS-ready listings, using prompting, RAG and structured tagging.
05Company
My company, Algeevra.
Through Algeevra, I help companies find where AI saves their team time, build it on their own data, and train their people to run it. Private by default, within EU rules.
01 AI Opportunity Audit
Find where AI pays off.
We find the tasks where AI saves your team the most time.
02 Tailored AI Education
Get your teams ready.
Hands-on workshops that teach your people to use AI safely and well.
03 Strategic Implementation
Build what matters.
We build AI that does real work inside your company.
04 EU AI Act Advisory
Stay compliant.
The EU AI Act, translated into plain terms for your management.
06About
Boardroom to terminal.
I’m an AI engineer and advisor who learns by building. I’m drawn to difficult problems, especially where AI can make a business work smarter.
I want the result to be useful, understandable and worth adopting. I keep learning because the field moves quickly, and because I enjoy going deep into a new technical challenge and finding the practical path through it.
On the business side, that means AI strategy, advising management teams, process analysis and AI education. On the technical side: model fine-tuning, multimodal AI, system evaluation, production integration and full-stack applications.
- Core tools
- Python, TypeScript, React, Node.js, FastAPI, PostgreSQL, LangGraph, PyTorch, OpenCV, Docker
- Education
- BASc Artificial Intelligence, Howest University of Applied Sciences
MSc Business Management & IT, Open Universiteit (expected February 2027) - Open source
- Contributor to Open WebUI, LiteLLM and AMD GAIA
- Outside work
- Quantum computing, the stock market, guitar and podcasts. Twelve years of basketball.
07Contact
Let’s build something that holds up.
Have a project in mind, a role to fill, or a question about AI? Send me a message. For company engagements, you can also go through Algeevra (opens in a new tab).