Services
Six disciplines, one standard.
We take on work in these six areas. Engagements usually cross more than one of them, because the problems do.
Data analytics consulting
Most organisations are not short of data. They are short of agreement about what it means. We build the definitions, the pipelines and the reporting that let a leadership team argue about the decision rather than about whose number is right.
What you get
- A defined metric layer — one agreed definition per measure, documented and version-controlled
- Pipelines that land clean, tested data on a schedule you can rely on
- Reporting built for the decisions it supports, not for decoration
- Analysis of the specific questions currently holding a decision up
When you’d call us
- Two departments report the same metric and get different answers
- Monthly reporting is assembled by hand and nobody quite trusts it
- You are about to buy a data platform and want to know whether you need one
AI engineering
Building something that works in a demo is straightforward. Building something that behaves predictably against your real data, at your real volumes, with known failure modes, is the actual job.
What you get
- Retrieval systems grounded in your own documents and records
- Agents and workflows scoped to tasks they can complete reliably
- Generative UI dashboards built with CopilotKit, where the interface assembles itself around the question being asked
- An evaluation harness, so changes are measured rather than guessed at
- Monitoring that surfaces degradation before your users report it
When you’d call us
- A prototype works and you need to know what running it in production costs
- A model gives plausible answers and you have no way to tell which are wrong
- You need AI reading internal data without that data leaving your control
Software engineering
Applications shaped around how your people already work. We spend time on the process before proposing software, because most failed internal tools fail for encoding a workflow nobody actually follows.
What you get
- Web, mobile and internal applications, built and deployed
- Command-line tools for operating and interrogating your systems, for the people who would rather not click
- Integration with the systems you already run
- Automated tests and a deployment pipeline your team can operate
- Documentation written for whoever maintains this after us
When you’d call us
- A critical process runs on a spreadsheet only one person understands
- Off-the-shelf software nearly fits, and the gap is expensive
- An existing application has become too fragile to change safely
Systems design & development
Architecture decided early determines what is possible later. We design for the load a system will actually carry, and state the trade-offs being accepted rather than letting you discover them under pressure.
What you get
- Architecture and data models, documented and reviewed with your team
- Licence servers and entitlement systems for software you distribute
- Integration design across the services and third parties you depend on
- Infrastructure, environments and deployment automation
- A written account of the failure modes and how the system behaves in each
When you’d call us
- Something that worked at ten users is struggling at a thousand
- Several systems need to agree with each other and currently do not
- You are planning a rebuild and want the design reviewed before committing
AI strategy & integration
A sequenced plan for where AI belongs in your operations, and a clear account of where it does not. The most valuable output of this work is often a shorter list than the one you started with.
What you get
- An assessment of candidate use cases, ranked by value and by feasibility
- A sequenced roadmap with the dependencies made explicit
- Data readiness and governance review, including what has to change first
- Integration design connecting AI to the systems where the work happens
When you’d call us
- Your board has asked for an AI strategy and you want one grounded in operations
- Pilots keep succeeding individually and never reach production
- You need an independent read on a vendor proposal
AI training & enablement
These systems only hold up if the people around them can judge their output. We train teams to brief, evaluate and supervise AI — the difference between a team that uses AI and a team that can tell when it is wrong.
What you get
- Practical training for the teams who will use these systems daily
- Evaluation rubrics your people can apply to real output
- Working practices and review processes fitted to your organisation
- Handover, so your team can extend the work without us
When you’d call us
- Staff are using AI tools informally and nobody is checking the output
- You are deploying AI into a regulated or high-consequence process
- You want internal capability rather than a permanent dependency on a supplier
What we build with
A deliberately narrow stack.
We would rather know five technologies deeply than twenty superficially. Where a project genuinely needs something else, we say so rather than reaching for the tool we already own.
- PostgreSQL
- The default datastore. Relational until there is a measured reason not to be.
- Rust
- Where correctness and throughput matter more than how fast we can write it — services, CLI tools and licence servers.
- Python
- Data work, analysis and the AI and evaluation pipelines.
- TypeScript
- Application and API code, typed end to end so the compiler catches what review would otherwise have to.
- React
- Interfaces, including generative UI dashboards built on CopilotKit.
Not sure which of these you need?
That is a normal place to start. Describe the problem and we will tell you which of these it actually is — including when the answer is none of them.