Self-hosted LLM deployment
CoreModel selection, GPU and CPU serving, quantisation choices, load handling, and the ops work to keep inference available under real traffic.
services
Everything between bare metal and a shipped product.
We take engagements at any point on that line — a one-off deployment, an ongoing maintenance relationship, or the whole build. Most clients start with one and grow into another.
Engagement
Project build
Engagement
Ongoing maintenance
Engagement
Embedded with your team
Engagement
Advisory / review
infrastructure
The layer everything
else depends on
else depends on
Infrastructure and platform operations
We provision, secure, and maintain Linux environments — primarily Ubuntu — and stay on as the people who pick up the phone when something breaks at 2am.
Deployment
- Server provisioning and configuration
- Service management with systemd
- Containerised and bare-metal workloads
- Reverse proxies, TLS, DNS, routing
- Environment separation: dev, staging, production
Maintenance
- Patching and upgrade planning
- Monitoring, logging, and alerting
- Backups plus tested restores
- Hardening, access control, key management
- Capacity and cost review
ai-ml
Models that run
where you decide
where you decide
AI and machine learning
We run multiple language models on our own infrastructure every day. That's the same service we offer you — including the part where we tell you a model isn't the answer.
Retrieval and data pipelines
CoreGetting your documents, records, or product data into a form a model can use — with the ingestion, chunking, indexing, and refresh cycle spelled out.
Evaluation and guardrails
CoreTest sets that reflect your actual users, measured before and after changes, plus the safety boundaries a system needs before it talks to the public.
Applied ML and automation
CoreClassification, forecasting, extraction, and workflow automation where a smaller purpose-built model beats a large general one.
product
Software people
actually use
actually use
Product and application engineering
Web applications, APIs, and open-source contributions — built with the deployment story figured out from day one, because we're the ones who'll be deploying it.
Web
Applications and sites
Front-end and back-end development, accessible interfaces, and performance that holds up on a mid-range phone.
API
Services and integrations
APIs, third-party integrations, background jobs, and the plumbing between systems that were never meant to talk.
OSS
Open-source work
Hosting, maintaining, and extending open-source software — including community platforms we run in production.
healthcare
Our chosen
specialism
specialism
Healthcare and behavioral health
We build in this space by choice, and we bring the habits it demands: careful data handling, a documented perimeter, and a bias toward keeping sensitive information on infrastructure you control.
- Privacy-first architecture and data-flow mapping
- On-premise and private-cloud model hosting
- Access control, audit trails, encryption at rest and in transit
- Product design for clinical and self-management contexts
- Working alongside your compliance and clinical advisors
- Plain documentation your team can hand to a reviewer
We're engineers, not a compliance authority or a clinical provider. We build systems to support your regulatory and clinical requirements and work with the people who own them.
who-we-work-with
Not limited
to one sector
to one sector
If it runs on a server, we can help.
Healthcare is where we're investing long-term. It isn't a filter on who we take calls from.
Early-stage teams
You have a product and no infrastructure person. We become that, until you hire one.
Established organisations
You have systems that work and nobody left who set them up. We document, stabilise, and modernise.
Teams exploring AI
You want to know what's real before committing budget. We'll tell you, including when the answer is "not yet".
get-in-touch