ManpowerGroup is looking for a Jr Architect with REVIT (H/M/X).
ManpowerGroup is a recruitment giant, one of those titans that for decades have connected companies with “human resources.” They are like an old, reliable, but slightly dusty bridge between those who have a task and those who can perform it. Now, they are looking for an architect for a major premium real estate developer in Spain.
Let’s see who they are looking for. They need a specialist to create luxury real estate. A person who will translate beautiful ideas into boring but mandatory drawings and BIM models in Revit. Project development for obtaining licenses, elaboration of structural details, 2D and 3D modeling, as well as author’s supervision on site. Special emphasis — they need a “colegiado,” meaning a certified specialist with a license, as they will be signing off on projects. Essentially, the company is looking for hands to do the technical work and a head to bear the legal responsibility for it. The pain point here is classic: the long, expensive, and error-prone process of manual design.
Now, let’s imagine we are solving not the problem of “finding a person,” but the problem of “creating a project.” And it turns out that for 90% of this work, a human in their classic form is the most inefficient link. A legal signature — that’s what they truly need from a living employee. Everything else is work for algorithms.
Allow me to paint you a picture. Let’s imagine a dialogue between a construction director (let’s call him Miguel) and an IT consultant (let’s say that’s me).
Miguel: We need another architect. Señor Gomez is bogged down in details for the new hotel, and we need to launch the residential complex in Marbella. Renders are ready, but working documentation is an eternity away. Manpower is already searching, but there are few competent ones, and they cost an arm and a leg.
Me: Miguel, are you sure you need a person? You’re not looking for an architect, but a Revit operator with a license. You pay them for 8 hours a day, 6 of which they will spend moving walls in a model, checking specifications, and fixing clashes. A machine can do this work overnight.
Miguel: A machine? These are serious projects, not a game of building blocks. There’s responsibility, norms, SNiPs… sorry, local building codes. Who will make the decisions?
Me: And this is where it gets interesting. You and your chief architect will make the decisions. But you will make them at a different level. Instead of saying “move this wall 20 centimeters,” you will give the machine parameters: “Design 10 layout options for this floor for me with maximum usable area, adhering to insolation norms, fire safety norms, and a budget of N euros per square meter.”
How does this work in practice?
Step 1: Generative Design. Instead of one architect spending weeks drawing a single option, we use tools like Autodesk Generative Design or Spacemaker. You set goals (view from the window, cost, energy efficiency) and constraints (legislation, plot boundaries). AI calculates thousands of possible options in a few hours and offers you 5-10 optimal ones. Your chief architect transforms from a creative drafter into a strategist who chooses the best concept.
Step 2: Automated BIM Modeling. After selecting a concept, platforms like TestFit or custom scripts for Revit (Dynamo) come into play. They take the approved scheme and automatically develop it into a detailed BIM model. Walls, floors, windows, basic engineering networks — all of this is created algorithmically, without “human error.” A process that takes weeks is compressed into days.
Step 3: Automated Checking and Documentation. Your new AI “employee” runs the created model through clash detection systems (intersections of pipes with walls) and, most importantly, for compliance with building codes and regulations (plugins and services like UpCodes exist). It also generates 90% of the necessary documentation for obtaining licenses: specifications, floor plans, sections.
How to overcome distrust? Start with a pilot project. Take an already completed object and assign the AI pipeline to replicate it. Compare the time, costs, and number of errors. The numbers will surprise you. The role of a human changes: they are no longer the “hands,” they are the “brain” and the “validator.” You don’t fire all architects. You hire one, but a very experienced and expensive one, for the role of controller and holder of that very license. They don’t draw; they check the machine’s work, make strategic adjustments, and apply their precious signature. Their time is spent on what AI cannot do: negotiations, aesthetics, final approval.
And how to validate the results of AI’s work?
Firstly, that senior architect-controller. Their experience is the best filter. But unlike checking the work of a junior employee, checking AI’s work is simpler. The machine’s logic is transparent; all parameters are fixed.
Secondly, simulations. Before building, run the digital model through simulators for load, energy efficiency, and airflow (CFD analysis). You can give the building a digital crash test even before the foundation pit is dug.
Thirdly, cross-checking. Ask the AI to generate an estimate based on its own model. If the numbers match the specifications, then the model is consistent.
Ultimately, instead of searching the market for another “Revit operator,” the company could invest in a system that does 90% of the routine work. And with the money saved, hire not just a specialist, but a true architectural star on a part-time basis — exclusively for control, creativity, and that very signature. But for this, one must stop thinking in terms of “vacancies” and start thinking in terms of “tasks.”
Источник: https://www.linkedin.com/jobs/view/4402890967/
