Three systems, described plainly: what the problem was, what we shipped, and which technology did the work. Client names are left out until we have written permission to use them. We would rather be dull than indiscreet.
Case summaries
High-volume applicant intake was being read by hand. We built the intake, the text parsing and a gradient-boosted scoring model, then put a review dashboard in front of the borderline band so a human still makes the call that matters.
A multi-tier platform with an Angular front end over a Django REST API: token authentication handled in interceptors, catalogues filtered in real time, and operational analytics for the team running it day to day.
An interactive product-discovery application for an FMCG brand, taken from concept to a deployed site: a headless content model the brand team edits themselves, served from the edge.
Longer technical write-ups live in Insights.
Get started
Describe the situation in a few sentences. We will tell you whether it is work we should take and what we would do first.
