AutoML Desktop App
Structured training runs for K-Means, Decision Tree and KNN — without hand-rolling another notebook. Every experiment is validated, logged and comparable.
In production · internal research tool · Python + Tkinter + scikit-learn
Why it exists
The setup cost more than the modelling
Each new question meant loading data, splitting the set, training a model, printing a score and then losing track of the experiment. The modelling itself was not the bottleneck.
Datasets were loaded inconsistently, parameters were not logged and comparing runs relied on memory. The app removes that repetition so proper modelling becomes the convenient choice.
What we built
Repeatable experiments by default
- Structured training runs for K-Means, Decision Tree and KNN
- Dataset loading and validation up front, before any model sees it
- Every run logged with its parameters, so experiments are reproducible
- Side-by-side model comparison against a baseline
- Results export for reporting and handover
How it works
A controlled run every time
Load a dataset
Select the file; validation runs before training begins.
Pick a model
Choose K-Means, Decision Tree or KNN.
Run
Parameters and split are logged automatically for reproduction or comparison.
Compare
Results appear side by side against the baseline, so the best result is visible directly.
Honest note
Measured internally before promised externally
The ≈80% figure reflects our internal experiment runs, not a client claim. We log numbers like this ourselves before promising them to anyone else.
Related service
Workflow Automation
We build internal tools around repetitive technical work as well as business operations.
View serviceFree automation audit
Want this kind of internal tool for your team?
Show us the repeated setup, manual comparison or reporting work. We will tell you what is worth automating and what it should cost.
No obligation · No sales deck · We reply within one business day.