Skip to content
Wide shot of the Industrial AI Lab

Industrial AI Lab

The Industrial AI Lab

Not a demo room. A working lab where problem statements from plants and MSMEs become trained models, tested on industrial hardware.

Our facilities

What's in the lab

Everything shot on our own bench — hardware our learners and pilot teams use every day.

Industrial cameras & lighting

Industrial cameras & lighting

GPU training server

GPU training server

IoT sensors

IoT sensors

Learner workstations

Learner workstations

On the bench

What we're working on

Each card shows the problem, the approach and where it stands.

Surface defect detection on finished product — public-dataset detection
Building

Surface defect detection on finished product

ProblemSpotting scratches, dents and surface flaws on finished product before dispatch.

ApproachIndustrial camera and lighting rig with an object-detection model, trained and evaluated on public datasets.

PPE compliance in work zones — public-dataset detection
Testing

PPE compliance in work zones

ProblemChecking that helmets, vests and other PPE are worn in marked work zones.

ApproachPerson and PPE detection on camera feeds, tuned for plant lighting conditions.

Material counting on conveyors — public-dataset detection
Building

Material counting on conveyors

ProblemCounting items moving on a conveyor without manual tallying.

ApproachDetection and tracking pipeline running on an edge AI device.

Reading analogue gauges from images — public-dataset detection
Scoping

Reading analogue gauges from images

ProblemReading needle-type gauges from photos to reduce manual rounds.

ApproachGauge localisation followed by needle-angle estimation.

Problem statements are generalised. No organisation is identified.

Pilot process

How industry works with the lab

  1. 1

    Share the problem

    A short call to understand the operation and its constraints.

  2. 2

    Scope it

    Agree the success measures, data needs and timeline.

  3. 3

    Pilot in the lab

    Build and test on representative data.

  4. 4

    Review results

    Measured against the agreed success measures.

  5. 5

    Decide next steps

    Scale up, refine, or stop, based on evidence.

Every engagement is handled independently. Data shared for a pilot is used only for that pilot.

Who it serves

Who uses the lab

Learners

Learners

Flagship programmes, offline and online.

Learn more
MSMEs and industries

MSMEs and industries

Pilots and proofs-of-concept.

Learn more
Institutions

Institutions

Faculty and student projects.

Learn more

Facilities

Built for learning at scale

Two connected smart classrooms
80+ learners

Two connected smart classrooms

Two connected smart classrooms for cohorts of 80+ learners.

Recording & podcast studio
Courses, lectures and industry content

Recording & podcast studio

A recording and podcast studio for courses, lectures and industry content.

The doorway of the Industrial AI Lab

See the lab for yourself.

Book a visit and see the bench, the hardware and the work in progress.