Benjy Bodner

When training a model, which “statistically righteous” approach do you instinctively trust the most??
Vote like your future ML career depends on it. (It doesn’t…:)

1 week ago | [YT] | 3

Benjy Bodner

Test your Gaussian skills!!

When computing the KL divergence between two univariate Gaussians N(mu1, sigma1^2) and N(mu2, sigma2^2), which term is always present in the closed-form expression?

1 week ago | [YT] | 2

Benjy Bodner

Many people mix up with the confusion matrix — can you get this right on the first try?

In a binary classification problem, which element of the confusion matrix represents the number of incorrectly predicted positives?

3 weeks ago | [YT] | 2

Benjy Bodner

The two opponents walk into the ring, each claims to have the upper hand. The data scientist pulls out a silver ruler, the deep learning developer pulls out a gleaming hammer — who will build the best model?

I put this classic debate to the test on a binary classification problem (regular customer vs. criminal) and the results are shocking:


Model centric approaches failed. Our simple and complex models (MLP, KAN, RF) were only as good as a coin toss on the initial data.
Data centric approaches won. By simply reversing an anonymization transformation (feature engineering), our worst model's performance improved by up to 8×!
Data first always wins. 🏆


👉 Read the full analysis on Medium: medium.com/data-science/deep-learning-vs-data-scie…

#datascience #deeplearning #datafirst #machinelearning #featureengineering #ai #pythonml

Skip the Carousel

1 2

1 month ago (edited) | [YT] | 2

Benjy Bodner

🧠 Quiz: Gini Impurity — The Metric That Trips Up 80% of Candidates!

In decision tree algorithms, what does the Gini Impurity measure?

1 month ago | [YT] | 2

Benjy Bodner

The Jaccard Index compares two sets by looking at their overlap and union. Which of the following metrics or formulas is identical in concept to Jaccard?

1 month ago | [YT] | 4

Benjy Bodner

We post deep-dive articles, career insights, and technical guides on Medium that perfectly complement the videos here on the channel.

Don't miss out on what the community is reading right now!

➕Top 12 Skills Data Scientists Need to Succeed in 2025

➕Deep Learning vs Data Science: Who Will Win?

➕What is more important: your data or your model?


➡️ Visit the Medium Page Here: medium.com/@benjybo7

#DataScience #DeepLearning #AI #MachineLearning #CareerAdvice #BenjyBodner

Skip the Carousel

1 2 3

1 month ago | [YT] | 5

Benjy Bodner

Which model can NOT be used for unsupervised clustering

5 months ago | [YT] | 3

Benjy Bodner

Which model is NOT typically differentiable?

5 months ago | [YT] | 2

Benjy Bodner

Did you ever implement a CNN for a project you worked on?

5 months ago | [YT] | 2