One problem, three distances from it. Learn to do it, have it done with you, or run the software yourself. Most people only ever need the first.
AI has not closed the skills gap in 3D. It widened it. The tooling moved faster than the people who can judge it, and the parts that decide whether a project succeeds still do not automate: what to capture, what to trust, what to ship, and when a pipeline is lying to you. Somebody has to hold that judgement. It may as well be your team.
Sometimes there is no time to learn it first. A deliverable has a date on it and it has to be right the first time. Other times the team is perfectly capable and what is missing is direction: somebody who has already made the expensive mistakes sitting in on the calls that are hard to reverse.
After enough client projects the same problems keep turning up in different industries wearing different names. Registration, segmentation, meshing, measuring, and exporting something a non-specialist can open. At some point building it once, properly, beats building it again for the next client.
Organisations whose teams I have worked with, through training, R&D guidance and software projects.
What each engagement involved stays between the client and me. What follows instead is the shape of the problems that keep coming back, and what the work actually produces. Every image on this page is output from my own tools, never client data.
Not slides. These are real results from the software, the same kind of output the courses teach you to build and the engagements deliver.
Every point labelled by what it belongs to: floor, walls, tables, chairs. The scene stops being something you look at and becomes something you can query.
Each object pulled out as its own instance across a full outdoor capture, so they can be counted, measured and exported one at a time.
Millions of points sorted by surface type across a whole district in a single pass, on hardware you already own.
A textured mesh rebuilt from aerial capture, light enough to open on an ordinary laptop rather than a workstation.
A messy scan reduced to clean planar geometry: the input a BIM or IFC export actually needs, instead of a billion loose points.
A 2D plan carrying real dimensions, generated from the cloud and traceable back to the points behind every number on it.
From holding the tape measure to shipping the software. Hover any underlined term for what it actually means.
Started on archaeological excavations, measuring walls with tape and cameras, then learning photogrammetry from the ground up. Two years of knowing exactly how much work sits behind every clean dataset.
A PhD in 3D geospatial sciences, introducing the smart point cloud, and the ISPRS Jack Dangermond Award along the way. Over 50 peer-reviewed papers, and a habit of reading the method section before the marketing.
Leading R&D for large organisations and advising startups, which is where you learn that the hard part is rarely the algorithm. It is the constraint nobody wrote down.
The Academy, and the software that came out of doing the same work often enough to productise it. Thousands of engineers trained, and a software suite that runs on their own machines rather than mine.
Six years ago I watched a senior engineer, brilliant with code, freeze in front of executives. The question was simple. Should we use this technology, or that one?
The Academy exists to build technical leaders who design systems instead of following tutorials. People who can walk into a room full of conflicting data, tight budgets and unclear tech, and still make the call.
Because in the AI era everyone can run the tools. The edge belongs to whoever knows which one to pick.
They connect, but none of them is a prerequisite for another. Pick the one that matches where you actually are.
Three questions, a 24-day roadmap and five missions. No card, no catch.
Every standalone course I publish, for as long as you subscribe. EUR 297 a quarter or EUR 897 a year.
The Library plus everything I do not sell separately. Three tiers from EUR 1,997, lifetime access.
Team training, R&D guidance, or co-development. Quoted after a 15 minute call.
The desktop engine. One file, no account, running entirely on your own machine.
Tools change every year. The pipeline does not. I teach the full chain, from capture to deployed insight, the way I wish someone had taught me. That is how you stop being dependent on whichever vendor is hot this quarter.
Dr. Florent PouxFounder, 3D Geodata AcademyEvery concept lands in a real project. No isolated theory, and no notebook that only works on the demo dataset.
You learn faster beside peers who are one step ahead of you, and one step behind.
When the field moves, the material moves with it. I add new methods as they land, not on a release schedule.
Transform raw 3D point clouds into intelligent spatial systems. From foundational concepts to production-ready AI models, this comprehensive guide takes you on a journey through the
complete 3D data science ecosystem, proven methods used by industry leaders at Netflix, Meta, and Airbus.
Also available on Amazon, Apple Books, and other major retailers
Perspectives from clients, readers and students who have worked through the material with me.
“The free mission showed me I did not need a PhD. I went from complete beginner to processing LiDAR scans in production within three months.”
“Used the code pack to pitch my CTO. Got approved for the full program plus the cloud budget.”
“Finally understood the full picture. The framework made everything click. I enrolled in the full program the next day.”
3,400+ engineers began exactly here. Three questions, a roadmap and your first working result. If you would rather talk it through first, that door is open too.
No credit card. No drip funnel. Unsubscribe whenever you like.
