Abhishek Saxena
Systems/03 · Service Intelligence
III03 / 08

Service Intelligence

A streaming computer‑vision loop — detection, tracking, state machines — that turns a restaurant floor into proactive service.

Hospitality group2024 – 2025Case studyDocs openDemo by arrangement
RoleArchitect
Built forHospitality group
StackYOLO · BoT‑SORT · TensorRT · Redis Streams · DVC · GPU pools
What is realReal pipeline shape
DataSynthetic floor footage

Walkthrough

01 · public
00:0004:20

A guided walkthrough of the Service Intelligence 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.

Chapters
  • 00:00The floor, three cameras
  • 00:50Detection and tracking
  • 02:00Per‑table state machine
  • 03:10Alerts, and why not on a single frame
  • 04:20Staff feedback to retraining

Live demo

03 · by arrangement
Signed out

Run the Service Intelligence demo.

A downscaled reimplementation of the production system on synthetic data, with a console and an inspector. Credentials are issued personally and expire.

What you will be able to do
  • iWatch tables change state in real time
  • iiRaise and clear an attention alert
  • iiiFlag a wrong alert as staff
  • ivSee the flag land in the retraining set

What is real here

04 · reality contract
Architecture

Real pipeline shape. The design, models and control flow are the ones that ran in production, rebuilt at a scale one machine can run.

Data

Synthetic floor footage. No customer data anywhere; generated sets that reproduce the real distributions.

Integrations

Cameras simulated. Stand‑ins are labelled as stand‑ins, in the demo and in the docs.