Personalisation & Recommendation
Behavioural signals to ranked results: recommendations built for sparsity, cold starts and business rules.
Walkthrough
01 · publicA guided walkthrough of the Personalisation & Recommendation demo on synthetic data: how the flow runs end to end and how the console is operated. Recording in progress; the chapters outline what it covers.
- 00:00Signals
- 01:00Ranking under sparsity
- 02:10Rules after the model
Documentation
02 · publicBusiness Problem & Solution
Why enterprise support needs grounded, cited, confidence‑scored answers — and what this downscaled demo proves.
User Journey
From a question to a cited answer — or an honest escalation to the right human team.
Architecture Document
The pipeline stages, the cost lever, the dual cloud/open‑source approach, and what is out of scope.
Presentation
The executive deck: problem, approach, results, what would change at full scale.
Live demo
03 · by arrangementRun the Personalisation demo.
A downscaled reimplementation of the production system on synthetic data, with a console and an inspector. Credentials are issued personally and expire.
- iBrowse as a cold‑start guest
- iiWatch the ranking change with behaviour
- iiiToggle a business rule
What is real here
04 · reality contractReal architecture. The design, models and control flow are the ones that ran in production, rebuilt at a scale one machine can run.
Synthetic behaviour logs. No customer data anywhere; generated sets that reproduce the real distributions.
Catalogue feed simulated. Stand‑ins are labelled as stand‑ins, in the demo and in the docs.