Open-Source · Technologies
Linux
The operating system your AI stack stands on.
Linux runs almost everything that matters: the servers, the containers, most of the cloud and nearly every serious AI deployment. When you self host AI, Linux is the ground floor. We build, harden and run Linux infrastructure for AI workloads, on your hardware or your cloud, and train your team to own it.
SERVER FLEET · EXAMPLE ESTATE
HOST
ROLE
PATCHING
STATUS
ai-gpu-01
Model serving
Current
HEALTHY
agents-01
Agent stacks
Current
HEALTHY
apps-01
Internal tools
Current
HEALTHY
backup-01
Snapshots
Due
REVIEW
legacy-box
Unknown owner
Never
RISK
Sample estate
healthy · review · risk
In plain terms
Free to download. Expensive to run badly.
Linux gives you control of the whole stack, if it is set up and kept properly. Here is what changes.
Without the work
- Servers built by hand, each one slightly different
- Patching that happens when someone remembers
- Security settings copied from an old forum post
- One person who knows how it all works, until they leave
With the work
- Servers built from a standard, repeatable configuration
- Patching and updates on a schedule, not a memory
- Hardening applied deliberately, to a baseline
- Documentation and skills so the estate outlives any one person
What CG TECH can do with Linux
The work, broken into the parts that matter.
The ground floor, done properly
GPU drivers, storage layout, networking and remote access set up correctly for model serving and agent stacks, because AI workloads punish sloppy foundations.
foundations that hold under AI
Security as a baseline, not a bolt on
Firewalls, access control, minimal services and disciplined remote access, so a self hosted stack is not a soft target on your network.
a stack that is not a soft target
Servers as a recipe, not a craft
Standard configurations mean the second server matches the first and the replacement matches the original, which is what makes an estate maintainable.
the second server matches the first
Owned by your team, eventually
We run the estate with you, document how it works and train your people, because infrastructure that only a consultant understands is a liability.
an estate your team can own
How an engagement runs
From a box in the corner to an estate, step by step.
01
Assess
We map what you run, on what, and where the risks sit.
02
Design
Standard builds, hardening baseline and backup strategy set.
03
Build
Servers rebuilt or brought to standard, documented as we go.
04
Handover
Patching rhythm, monitoring and skills your team owns.
Questions we hear a lot
Common questions about Linux
What is Linux?
Linux is an open source operating system. It runs most servers, nearly all containers, most of the cloud and almost every serious AI deployment. When you self host AI, Linux is the ground floor.
Why Linux for AI instead of Windows?
Because the AI ecosystem lives there. Model runtimes, GPU tooling and agent frameworks target Linux first, and most of the cloud you would otherwise rent is Linux underneath. It is the path of least friction and most control.
Is Linux really free?
The software is. The honest costs are hardware, power, time and skills. For steady workloads that trade often beats per seat and per token pricing, and we model it plainly before you commit.
Do we need Linux skills in house?
Eventually, yes, and we build them. Engagements include documentation and training, so the estate does not depend on us forever. That is deliberate.
What about support when something breaks?
Enterprise distributions offer paid support, and we run and support estates as part of an engagement. Between a supported distribution and a documented standard build, breakage becomes routine rather than a crisis.
Which version of Linux should we run?
For business AI work we normally stay with a long term support release from a major distribution. The support window runs for years and the driver support for AI hardware is well trodden. We pick it with your team rather than for them.
How do we keep it patched?
The same way you keep anything patched: a schedule, a test machine and a way back. We set that up as part of the build so it is not left to whoever remembers.
What if the person who knows Linux leaves?
That is the risk worth designing out. We document the build, keep it reproducible and train more than one person. Where you would rather not carry it at all, we can run it for you.
Ready when you are
Building on your own infrastructure? Let us talk.
A discovery session maps your workloads, your risks and your quick wins. You keep the plan either way.
What to expect
- A consultant replies within 4 business hours
- Session booked to understand your requirements
- We will provide you with a fixed price quote