📰 Medium · Machine Learning

A hands-on tour of Data Version Control, built on one small wine-quality model, from a frozen import to a cloud remote and a CI job Continue reading on Medium »

The model is just one component. The system is everything around it: context, tools, orchestration, evals, and cost controls. Continue reading on Medium »

From RFM segmentation to predictive Customer Lifetime Value using BG/NBD and Gamma-Gamma models. Continue reading on Medium »

Continue reading on Medium »

When people think about difficult machine learning problems, they often think about building more sophisticated models. Should we use a… Continue reading on Med

The gap between the demo and the deployment is where most AI projects die. This is what that gap actually looks like. Continue reading on Medium »

Artificial intelligence has moved beyond experimentation. Today, executives are under pressure to turn AI investments into measurable… Continue reading on Mediu

Claude Sonnet 5 vs Frontier models like Opus 4.8, GPT 5.6 series: a practical breakdown of pricing, tokenizer inflation, benchmark gaps… Continue reading on Dat

The moment “good enough AI” became good enough for production. Continue reading on Medium »

In the world of generative AI and LLM workflows, speed is UX. Continue reading on Medium »

Raw data is rarely ready for machine learning. Before training a model, we need to clean and transform the data so algorithms can… Continue reading on Medium »

Every day companies make thousands of hours of video from security cameras, production lines, warehouses and stores. Most of this video is… Continue reading on

Subtitle: Understanding the fatal flaws of linear Retrieve-and-Generate pipelines — and why the industry is moving to Agentic workflows. Continue reading on Med

We analyzed 170,927 AI research papers posted to arXiv since the beginning of 2025 up to June 26th across its four main machine-learning… Continue reading on Me

Traditional science and classical mathematics were built on the pillar of rigorous precision and categorical truths. However, when trying… Continue reading on M

The dirty secret of AI stock picking lives inside the model’s weights. Full write-up, code, and benchmarks on jiripik.com. Continue reading on Medium »

After spending the last few years building data pipelines with Python and Spark, I realized there was one area of the data ecosystem I had… Continue reading on

It trained on the world, then paid $1.5 billion to settle the pirated books. Its terms say you cannot use Claude to build a competitor… Continue reading on Towa

In this article, we’ll take a detailed look at how to pay for various purchases using a cold wallet, with a focus on Bitcoin transactions… Continue reading on M

After five years on the other side of the table, I finally met a candidate who made me rethink what good actually looks like. Continue reading on PracHub »

Most people think machine learning is mainly about choosing the best model. Continue reading on Medium »

Four teams, four products, zero coordination — and the same four ingredients show up in every one. Continue reading on Towards AI »

A few years ago, writing code felt like a very manual job. You opened your editor, typed everything yourself, searched Stack Overflow for… Continue reading on M

Ridge vs Lasso, and the One Picture That Ends the Argument Continue reading on Towards AI »