Two AI trainers, one shared idea
VectorCraft began with a conversation between two people who had been doing the same work from opposite sides of the screen.
How Alex and Nicholas met
In 2023, Alex Chen, a software engineer, was spending evenings evaluating model-generated code on AI training platforms. Nicholas Wang was doing the same kind of work, ranking responses and writing expert answers for reasoning projects. They met in an online community of AI trainers, where they kept answering the same hard questions about project guidelines. Before long they were reviewing each other's work.
The idea they shared
They had both noticed the same problem. Talented people were doing AI training alone, with no one to check their reasoning and no way to share what they learned. Quality depended on luck. Alex and Nicholas believed AI training should work like a good engineering team, with peer review, shared playbooks and people working in the subjects they know best.
“The models are only as good as the people teaching them. We wanted to build the team we wished we'd had when we started.”Alex Chen & Nicholas Wang, co-founders
Building the team
They started with two people and a shared document of guideline notes. They invited engineers, mathematicians and writers they trusted, set up a two-layer review process, and turned every piece of reviewer feedback into a lesson for the whole group. Approval rates went up, and so did the number of people who wanted to join.
From training models to shipping products
Working with models every day gave the team a deep feel for how they behave. Clients began asking whether we could build with AI as well as train it. We started an engineering practice and delivered our first full stack and AI-powered products: knowledge assistants, automation tools and web platforms that went live on time and kept working. Today VectorCraft does both, and each side makes the other better.



