Applied AI applications, reasoning workflows, product infrastructure, model design and customisation, and secure cloud model operations.
Client engagements / Selected workSystems we were
Systems we were
hired to build.
Alongside our own products we take on a small number of AI engineering engagements: applied AI applications, product infrastructure, model design and customisation. This work funds the product roadmap and keeps us building under real production pressure.
A small number of engagements at a time. We decline work that would pull the team off the product roadmap or outside what we build daily.
These systems belong to our clients. They are not Dovichi products, ventures or revenue lines we own, and we list them separately for that reason.
The brief
A UAE-licensed brokerage selling Dubai property to buyers across several countries needed one consistent intelligence layer behind a distributed sales team.
What we built
We structured property knowledge behind a Claude-based reasoning layer so the team answers complex buyer questions consistently and faster, with the same source of truth in every market.
ClaudeStructured product knowledgeSales workflowLead capture
The brief
A growing Dubai holiday-home portfolio was managing listings, availability, media and pricing across several distribution channels by hand.
What we built
We built a connected operations system that holds listings, availability, media and pricing in one coordinated workflow, cutting repetitive administration and keeping channels consistent.
Operations automationListing distributionProperty dataChannel sync
Our own products
The product line is where we are going.
Client work funds the roadmap. Our ventures — DoviLearn, LearningFilms and CVSense — are the business we are building, covering the path from learning a skill to being hired for it.
Next / Dovichi Labs