Work
Systems we’ve designed and built
For each project: the problem, what we built, and a simplified diagram of how it fits together. The diagrams are high level on purpose and leave out anything confidential.
AI revenue platform
RevAlign
A revenue platform that uses AI to suggest what sales and revenue teams should do next, based on territory data and the role of the person looking.
- The problem
- Turn day-to-day revenue data into recommendations that people in different roles can act on.
- What we built
- Recommendation models, territory analysis, operational analytics and role-based access, on an architecture built to grow.
- Outcome
- A clear product and platform direction, with recommendations shaped by how each team works day to day.
Related service: AI Engineering & Intelligent Automation, Custom Software & Product Engineering
- 1Revenue and operational data is turned into signals and features, including territory-level intelligence.
- 2Candidate recommendations are generated and ranked by AI models.
- 3Role and policy filters decide what each person sees, so a manager and a sales rep get different recommendations.
- 4Users act on recommendations in a role-aware app, and their responses flow back as feedback signals.
MVP platform architecture
OpenCompare
We designed, built and launched the first version of a platform for structured, side-by-side comparisons.
- The problem
- Get from concept to a usable platform quickly, without building more than the first version needed.
- What we built
- Backend foundations, the comparison workflows, and a data model designed around them.
- Outcome
- A launched MVP built around the core user journey, with an architecture that can take on new features as the product grows.
Related service: Custom Software & Product Engineering
- 1External data is ingested and normalised into a model where options can be compared on consistent attributes.
- 2Comparison logic applies rules and scoring on top of that data.
- 3A backend API exposes the comparison workflows to a UI built around structured user journeys.
Mapping infrastructure
Geo-Spatial Tile Infrastructure
Raster tile generation and geo-spatial infrastructure, built by our founder and used in production mapping workflows.
- The problem
- Turn large geo-data sets into map tiles reliably, and serve them at a consistent speed.
- What we built
- Geo-data processing, raster tile generation, tile delivery that scales with demand, and a caching strategy.
- Outcome
- Faster infrastructure, and mapping workflows that are easier to run and extend.
Related service: Cloud, DevOps & Platform Engineering
- 1Large geo-data sets are processed and prepared for rendering.
- 2Tile generation turns prepared data into raster tiles that are stored ahead of demand.
- 3A cache in front of tile storage keeps delivery fast and predictable for mapping workflows.
Want to hear more about any of these?
We’re happy to talk through the architecture decisions on a call, as far as client confidentiality allows.