Deployment and support
Hardware is half the job. Before a system reaches you it is burn-in tested under sustained load and the software you need is installed; at handover you get a working system and a record of how it is configured. Ongoing support is optional and scoped per project.
From parts to a working system
Assemble
Built to the agreed configuration, with cabling routed so it does not block airflow.
Firmware and BIOS
Updated to stable releases and set for the workload, including PCIe, memory and power settings.
Burn-in
Sustained GPU, CPU and memory load to surface early faults and thermal problems before delivery.
Software install
Operating system, GPU drivers, CUDA or ROCm, containers and the AI stack agreed in the quotation.
Verification
A representative workload, or one you provide, run on the finished system and its results recorded.
Handover
A walk-through, every account and credential, and the configuration record.
What handover includes
Configuration record
Components, serial numbers, firmware, driver and software versions.
Burn-in summary
What was tested, for how long, and the outcome.
Access
Administrator accounts and credentials, handed to you to change.
Operating notes
How to update, restart and check the system, and how to reach us.
Choose how much we do
Tell us which fits; the quotation is scoped to match.
| Option | We do | You do |
|---|---|---|
| Supply only | Build, burn-in and deliver the hardware | Install the software and deploy it |
| Setup only | Install and configure the software stack on hardware you already have, after a compatibility check | Supply and maintain the hardware |
| Full deployment | Build, burn-in, software, on-site installation and handover | Provide the facility, network and access we agree |
Software we can install
Common choices. Each is checked against your GPU, operating system and versions before it goes into a quotation.
Operating system
Linux, usually an Ubuntu LTS release, or Windows for workstation applications.
GPU platform
NVIDIA drivers and CUDA, or AMD drivers and ROCm, matched to the card.
Runtimes
Containers, and model runtimes such as Ollama, llama.cpp or vLLM, with a chat or API front end.
Frameworks
Python environments with PyTorch and the libraries your team uses.
Support after handover
Support is optional and scoped per project until standard terms are published. It can include remote help, software and driver updates, periodic health checks and on-site visits. Your quotation states exactly what is included.
How warranty and support workCommon questions
Can you set up software on hardware we already own?
Yes, after a compatibility check of the GPU, drivers and operating system against the software you want to run. Some older or consumer hardware will not support current AI stacks well, and we will tell you if so.
Who owns the administrator accounts?
You do. Every account and credential is handed over, and we recommend changing passwords at handover.
Can you install the system on site?
On-site installation is part of full deployment and scoped per project. Work outside Malaysia is confirmed per enquiry.
What if something fails after handover?
Report it and we work through it with you under the terms in your quotation. The process is on the warranty and support page.
Request deployment
Tell us what is being deployed, where, and which option fits. We scope the work and quote it.