Skip to content

Ollama Setup · Durham

Ollama Setup in Durham

Setting up Ollama for Durham teams who want models running locally, usually for development or for work that cannot leave the machine. What determines whether it is usable is memory: a model has to fit, and the quantised version that fits comfortably behaves differently from the full-precision one you may have evaluated.

Why this comes up

The problem

A model was chosen on published quality, the quantised version that actually fits produces noticeably worse output, and nobody compared them.

What you get

What we deliver

  • Model and quantisation selected against the memory actually available
  • Quality compared between quantisation levels on your own tasks
  • Hardware acceleration configured, since the difference is not marginal
  • Context length set deliberately, as it consumes memory alongside the model
  • A consistent local setup across the team rather than everyone improvising
  • An honest position on what this is for, and what it is not

Want this scoped for your business in Durham?

Thirty minutes, no charge, no sales script. You leave with a written summary of what ollama setup would actually involve — whether or not you use us.

Book a 30-minute call

Working in Durham

North East

Durham's economy is dominated by its university and cathedral, so the commercial base is small and much of the local demand comes from spinouts and suppliers attached to the institutions rather than from independent industry.

Research spinout software and visitor systems, usually on institutional procurement timelines rather than commercial ones.

Sectors we work with in Durham

  • Higher Education
  • Tourism & Heritage
  • Public Sector
  • Retail
  • Research

What we work with

Technologies and platforms

  • Ollama
  • Python
  • Docker
  • llama.cpp

Who we work with

Industries we serve

  • Technology companies
  • Research organisations
  • Healthcare
  • Legal services
  • Financial services

Why us

Why Durham businesses choose Asionis

  • Projects typically launched within 4–8 weeks
  • No long-term contracts required
  • All team members UK-based
  • Dedicated account manager and development team
  • Transparent reporting with monthly performance metrics
  • Scalable from startup to enterprise

How we work

  1. Step 1

    Free consultation

    A 30-minute call to understand the problem. You keep the written summary either way.

  2. Step 2

    Proposal

    Scope, timeline and a fixed price, in writing, before anything starts.

  3. Step 3

    Build

    Short cycles with regular check-ins, so you see progress rather than hear about it.

  4. Step 4

    Launch and support

    We handle the go-live and stay available afterwards.

Ollama Setup in Durham — common questions

How much does quantisation affect quality?
Enough to test. Heavier quantisation reduces memory considerably and degrades output, and how much depends on the model and the task — so the comparison should be run on your actual work rather than assumed from general reputation.
What limits which model we can run?
Memory, primarily, and it is shared with the context window. A model that loads with room to spare can still fail on a long prompt, so the sizing calculation has to include the context length you intend to use rather than just the weights.
Is local as good as a hosted model?
Generally not, at comparable convenience. Models you can run on a workstation are smaller than the frontier hosted ones, and that shows on harder tasks — which makes local the right choice when confidentiality or offline operation matters more than capability.
Should the whole team use the same setup?
Yes, or you will chase differences that are environmental. Different models, quantisations and context settings across developers produce inconsistent behaviour that gets attributed to code changes rather than to the setup.
Do you work with businesses across County Durham?
Yes. We are based in Leicester, United Kingdom and work with clients throughout North East, including Durham and the surrounding County Durham area. Most collaboration happens remotely, and we travel for kick-offs and key milestones.
What kind of Durham businesses do you usually work with?
Research spinout software and visitor systems, usually on institutional procurement timelines rather than commercial ones. Beyond that we work across Higher Education, Tourism & Heritage, Public Sector, Retail and Research.
Do you cover the areas around Durham?
Yes — we work throughout North East, including Newcastle, Sunderland, York. Durham is an urban area of roughly 70,000+, and we take on work across the wider County Durham region rather than the city boundary alone.

Talk to us about Ollama Setup in Durham

A 30-minute call with someone who would actually work on it. No sales script, no obligation.

  • Projects typically launched within 4–8 weeks
  • No long-term contracts required
  • All team members UK-based
  • Dedicated account manager and development team
  • Transparent reporting with monthly performance metrics
  • Scalable from startup to enterprise

Or reach us directly

07707 771599admin@asionis.com

Leicester, United Kingdom