Project-specific AI experts · Hardware included
In development
We adapt an existing AI model to your company, project, and data. We prepare its agents and deliver everything on a rented AMD-based AI computer. Your team works through a simple interface. We handle setup and support.
Planned service. Project rental, from setup to return.
01 / From your request to your desk
Explore the planned rental process
01 / Your starting point
Bring the project, the questions your team needs help with, and the expected rental period. Together, we define what useful expert support should do.
02 / We build your expert
We select an existing AI model, adapt the setup to your company and project, and prepare a knowledge base from the material you provide. We evaluate it against agreed project questions.
03 / More than a conversation
We configure task-specific agents to work through defined steps: find relevant context, plan the work, prepare outputs, and check them. Tools, permissions, and review points are agreed for the project.
04 / Hardware included
We deliver the model, knowledge base, agents, and interface together on a rented AMD-based AI computer. We handle setup and onboarding so your team can begin the agreed pilot.
05 / Your team uses it
Ask specialist questions, explore ideas, and delegate defined tasks through the interface. Review the supporting context, inspect prepared outputs, and approve actions where required.
06 / A defined finish
The setup is rented for an agreed period. At the end, return the computer and complete the agreed handover and removal of project data.
02 / Inside the computer
We build the setup around the company and the current project. The model provides reasoning, the knowledge base adds context, and configured agents help carry out defined work.
Adapted from an existing model for your project.
Company context, project files, and specialist material.
Instructions and tools for repeatable project tasks.
A place to ask, delegate, review, and continue.
03 / A teammate for the task
The planned agents can progress through agreed workflows with a defined level of autonomy. Your team sets the scope, reviews the result, and approves actions at the agreed points.
Open the interface demo“Prepare a work plan for this project.”
Find the objective, constraints, and relevant knowledge.
Draft tasks, dependencies, and open questions.
Organize the proposed plan and its assumptions.
Your team checks the plan before acting on it.
“Compare these options for our project.”
Read the requirements and material provided.
Compare the same factors across each option.
Explain differences, missing evidence, and uncertainty.
Your team reviews the reasoning and chooses.
“Check this proposal against our requirements.”
Retrieve the current project criteria.
Connect claims and sections to those criteria.
List unsupported assumptions and missing details.
Your team checks the findings and next steps.
Illustrative workflows. No live AI or autonomous actions run in this preview.
04 / Compact hardware. Local execution.
The planned setup uses compact AMD-based AI computers. We select the model and runtime for each project, then prepare the computer before delivery.
Reference platform: AMD Ryzen AI Halo with Ryzen AI Max+ 395 and 128GB unified memory. AMD lists support for models up to 200B parameters. View AMD’s hardware specification ↗
Planned AIEX target configuration
Up to
200B
parameter models
A project-adapted model, its knowledge base, and configured agents on a rented local computer.
Model size, format, context capacity, and speed must be validated for the selected hardware. AIEX has not yet benchmarked a 200B deployment.
05 / Built around your company
Local processing is the starting point for greater control over company data. Access, connected tools, retention, and review points are planned around your project’s requirements.
We prepare a project knowledge base from the approved material you supply.
Define which tools and actions belong in the workflow and which require review.
The planned service includes model adaptation, hardware setup, onboarding, and support.
Agree on the data handover, removal process, and return of the computer.
Founder-led in Düsseldorf, Germany. Discovery conversations are research, not customer endorsements. Pricing is being designed below the cost of a full-time specialist for the same project period.
Let’s build around your work
What is your team working on? What expert help does it need? We’ll scope the model, agents, hardware, and rental period around the project.
Explore the interface demo