Machine Learning & Computer Vision · Portfolio
We train prediction models and computer vision systems, test them honestly on unseen data, and deliver them as simple tools, reporting where they are strong and exactly where they fail.
For teams that want to predict, classify or detect something, and need to trust the answer.
01Selected work
The strongest projects first. Each opens a case study with the problem, what we built, screenshots and results.
MAUK R&D · applied machine learning
Risk scores for ten conditions from 25 routine measurements, trained on 281,000 records and private enough to run on the patient's own device.
86.88%accuracy on patient records the model never saw
Read the case study
MAUK R&D · computer vision
A road-scene detection model trained from scratch on 70,000 images with a laptop GPU, and an honest account of where it struggles.
0.575detection accuracy (mAP50) on 10,000 unseen images
Machine learning
LiveMAUK R&D · applied machine learning
A spam classifier built from first principles that explains every verdict it gives.
83.8%accuracy on 920 emails the model never saw
Machine learning
MAUK R&D · applied machine learning
From a student dataset to a working predictor: two models compared honestly, and the one that won shipped.
0.617fit on students the model never saw (R²)
Machine learning
MAUK showcase store for beauty & nail brands
Shoppers see the set on their own hand before they buy.
AI agents · E-commerce · Machine learning
LiveMAUK demo build · retail conversion scoring
A retail scoring interface that shows its reasoning: five sub-scores, one verdict, and the customer signals behind each.
Machine learning · Custom software
02Problems we solve
If any of these describe your week, the work above is the proof we can fix it.
You have data but no way to turn it into predictions
A single accuracy figure hides the mistakes that matter
Images or video still have to be checked by hand
Running a model costs more than it should
03What you get
Models that score risk, predict outcomes or classify messages.
Detection and tracking in images and video, tested in hard conditions.
Every prediction comes with the reasons behind it.
Accuracy reported on unseen data, broken down by the cases that matter.
Industries this work has shipped in
Prior work
Systems our engineers built in earlier roles. The clients are confidential, so there are no screenshots, and these are not presented as MAUK client projects.
Prior work · client confidential
Computer vision
The client's system is not shown
Prior work · national cricket body (confidential)
Computer vision that follows a small, fast ball across broadcast frames.
Machine learning
How we work
Technical detail on any project above is shared in a call, once we understand what you're building.
We learn your business and your constraints, and find where software or AI will pay off first.
We shape the solution with clear milestones and success measures, agreed before the build starts.
We build in short cycles and show working software as we go, so there are no surprises at the end.
We launch, watch how it is used, and keep improving it with you as the business grows.
Start a conversation
Tell us what you're trying to build. We'll reply with how we'd approach it, and which of the projects here are closest to it.
