Master's graduate · thesis completed 2026

About DataScientist

An advanced AI text detection platform built to support academic integrity.

Kungratov IlmurodMaster of Data Science (completed)Nordik International University

What is DataScientist?

DataScientist analyses text with ensemble machine learning to estimate how likely it is to be machine generated. Several complementary techniques are combined so the result does not depend on a single model.

Whether you are reviewing student submissions, verifying content as an editor, or researching academic integrity, the platform gives you the signals you need to make an informed decision.

How it works

Each analyser scores the text on its own. The final verdict is their weighted combination, which is far more reliable than any single method.

Perplexity analysis

Measures how predictable the text is. Machine generated text usually scores lower because it follows the patterns the model was trained on.

Burstiness detection

Analyses the variation in sentence length. Human writing is far more varied, while AI text tends towards uniformity.

Transformer classification

Uses a RoBERTa network trained specifically to recognise content written by GPT-family models.

Stylometric analysis

Examines vocabulary richness, word patterns and writing style markers that separate human from machine authors.

Models in use

The platform runs the following models:

desklib/ai-text-detector-v1.01

desklib/ai-text-detector-v1.01 — a DeBERTa-v3-large classifier trained on the RAID benchmark, used for document- and sentence-level evidence. It carries the largest share of the ensemble weight.

distilgpt2

distilgpt2 — used for perplexity and entropy analysis. It detects the low-perplexity signature of AI text.

Burstiness

Burstiness — measures how much sentence length and complexity vary. Human writing is markedly more 'bursty'.

Stylometry

Stylometry — uses NLTK for POS tagging, lexical diversity and AI-phrase detection.

Ensemble

Ensemble — the results of all four are combined using calibrated weight coefficients to produce the final AI probability.

An important note

DataScientist produces probabilistic assessments and should be treated as one signal among several. No AI detection tool can guarantee perfect accuracy, and the results must never serve as the sole evidence of AI authorship.

Try DataScientist